Shinji Yamasaki* and Ken Takahashi**
*Doctoral course student, Department of Archaeology,
Graduate School of Humanities and Sociology, the University of Tokyo)
**Research fellow at the Japan Society for the Promotion of Science;
Department of Archaeology, Graduate School of Humanities and Sociology,the University of Tokyo)
This monograph reports on the archaeological findings from the Hikosaki Shellmound owned by the Department of Physical Anthropology and Prehistory, the University Museum, the University of Tokyo. These materials were excavated at the Hikosaki Shellmound by the Anthropological Laboratory of the University of Tokyo in 1948 and 1949. The materials, contained in 40 wooden cases, consist of a large number of sherds, stone implements, bone artifacts, and faunal remains. There are also a large number of human skeletons, probably belonging to the Early and Late Jomon period; these skeletons have already been catalogued and are stored in the National Science Museum (Endo and Endo 1979).
In 1950, this collection was studied by Touju Ikehasu, a member of the excavation team at the Hikosaki site, who later published a brief report (Ikehasu 1971). The Hikosaki Shellmound is well known as the type site for the Hikosaki type pottery. Sugao Yamanouchi, a famous archaeologist who constructed the framework of Jomon chronology, distinguished four pottery types based on the Hikosaki collection: Hikosaki Z1, Z2, K1, and K2 (Z stands for zenki, or the Early Jomon, and K stands for kouki, or the Late Jomon). However, Yamanouchi did not publish the details of his study. Consequently, while the Hikosaki type became very important in the chronology of western Japan, its detailed nature remained unclear apart from the partial description by Kamaki (Kamaki and Kimura 1955; Kamaki and Takahashi 1965) and Ikehasu (1971).
With this as the context, our report primarily aims to describe the Hikosaki type pottery in detail. For this purpose, the excavated sherds were systematically observed, reclassified, drawn to scale, photographed, and catalogued. In addition, the stone implements and bone artifacts were also re-studied to facilitate comprehensive understanding of the site.
Shinji Yamasaki undertook the analysis of Jomon pottery and stone artifacts, and Ken Takahashi was responsible for the study of the artifacts made of bone, antler, tusk, and shell. Fumihiko Matsuyama offered valuable perspective on the stone tools. Norihiro Ueno took photographs of the objects. We would like to thank the graduate and undergraduate students of the University of Tokyo, including Masanori Oki, Tomoyasu Kiuchi, Takuya Saito, Masashi Taga, Shinji Tanaka, Kenji Nagai, Yuki Nakamura, Hiroyuki Hatsukano, Yoshihisa Furusawa, Kazutaka Morisaki, for their cooperation. We are also grateful to Toju Ikehasu, Takura Izumi, Noriko Ito, Keiji Imamura, Yosuke Kaihu, Reiko Kono, Tadashi Goto, Aiko Saso, Hiroyuki Sato, Gen Suwa, Masako Takahashi, Mamoru Takahashi, Masanori Tajima, Yutaka Chiba, Yoshihiro Nishiaki, Yasutami NishidaHisao Baba, Wakiko Noguchi, Tadahiko Makabe, Yoshiko Makabe, Soichiro Mizushima, Ken’ichi Yano, and Satoru Yamada, for the advice and accommodation support extended by them. This project was financially supported by the Takanashi Academic Fund. This monograph was authored by Yamasaki and Takahashi and was translated into English by Takahashi.
(Shinji Yamasaki, Ken Takahashi)
The research was executed at the University Museum, the University of Tokyo, from April 2002 to March 2005. We first recorded the original positions of the materials in the wooden cases. Following this, we alphabetically coded the packages in each case. After this process, the materials were marked, refitted, and drawn to scale. They were roughly divided into four categories: sherds, stone implements, artifacts made of bone, antler, tusk, and shell, and faunal remains. Most of the sherds had red markings indicating the excavated trenches and layers; for example, ‘1-4’ refers to Layer 4 of Trench 1. They were labelled and stored in wooden cases; as a rule, each case contained sherds from a single phase (Early, Middle, or Late Jomon) from a single trench.
The materials had been originally packed with newspapers, wrapping papers, or plastic bags.Table 1 shows the dates of the newspapers used for the packing. While there are newspapers from 1935 to 1957, newspapers from 1957 are most numerous. Since the volume of newspapers after 1950 exceeds that of newspapers before 1950, it appears that the research on the materials from the Hikosaki Shellmound was continued after the study by Ikehasu in 1950.
(Shinji Yamasaki)
The Hikosaki Shellmound is located on a small terrace approximately 6 m above sea level, facing the ancient coastline of the Kojima-wan Bay. The town of Nadasaki belonging to Okayama Prefecture (where the site is situated) was merged with Okayama City in 2005. Many shellmounds belonging to the Jomon period are located around the Hikosaki Shellmound, along the ancient coastline of the Kojima-wan Bay; for example, Satogi, Fukuda, Funamoto, and Isonomori are famous as type sites of Jomon pottery (Fig. 1 and 2). While the Hikosaki site revealed remains from the Initial to the Final Jomon period, it was mainly formed between Early to Late Jomon periods.
In 1947, the eastern half of the Hikosaki Shellmound was destroyed due to the expansion project undertaken by the neighbouring brewery. This incident triggered a series of excavations by the Anthropological Laboratory of the University of Tokyo. The excavations were executed thrice in 1948 and 1949, including the trial trenching on May 31, 1948 (Fig. 3). The trial trenching was conducted by Nakao Sakazume who had been interested in the Hikosaki Shellmound in the pre-World War II days and had thus observed the destruction of the site. During this trial excavation,he found remnants from Late and Early Jomon pottery from the upper layer and lower layer distinctively.More noteworthy were the uniquely styled burials unearthed below the shell layer; in a pit covered with ashes, Sakazume found three skulls placed on piled limb bones. In response to this remarkable discovery, the first extensive excavation was executed from 16 to 26 November 1948; the team opened five trenches (Trenches 1-5) between the trial trench and the edge of a cliff, which had appeared as a result of the destruction at 1947. The second excavation was executed from 3 to 17 August 1949; they opened nine trenches (Trenches 6-14) on both the sides of the previously excavated area. This series of excavations brought to light a large number of skeletons,sherds, stone and bone artifacts, faunal remains, and so on. Several brief reports exist on the findings from these excavations (Sakazume 1951, Ikehasu 1971, etc.)
(Shinji Yamasaki)
Ikehasu has already described in detail the overall stratigraphical situation of the Hikosaki Shellmound (Ikehasu 1971). In this section, we will summarize the existing information in order to understand the archaeological contexts of the materials.
The stratigraphical division at each trench was reconstructed from the markings, labels, and references of the materials (Table 2). It may be assumed that the layers measure approximately 20 cm in thickness on an average and were numbered in top-to-bottom order. In some cases, we could reconstruct the correspondence of the layer number with the substance of the layer, such as a shell layer or an intermediate soil layer. However, in most cases, especially during the second excavation, it was impossible to do so.
According to Ikehasu (1971), two shell layers were divided by an intermediate soil layer; while the lower shell layer (Shell Layer No. 1) contained Hikosaki Z1 type pottery, the upper shell layer (Shell Layer No. 2) contained pottery belonging to the Hikosaki Z2 type and other types from later periods. Shell Layer No. 1 was the thickest in Trenches 12 and 13, and became thinner towards the south. On the contrary, Shell Layer No. 2 was the thickest in Trenches 5, 10, and 11. In spite of our efforts, it was impossible to fully reconstruct the relation between the layers in the different trenches. We will later discuss the relation between the layers and the pottery types in
each trench.
(Shinji Yamasaki)
Table 3 shows the chronology currently in circulation, ranging from the Early to the Late Jomon period in the Setouchi region where the Hikosaki Shellmound is located. This chronology has been established through a series of excavations by Kamaki Yoshimasa and the Kurashiki Archaeological Museum, based on the pioneering study by Sugao Yamanouchi (Yamanouchi 1937).
The report on the excavation at the Satogi Shellmound, published by the Kurashiki archaeological museum (Makabe and Makabe 1971), occupies a particularly important place among the numerous studies existing on this subject, since it established the framework of Early to Middle Jomon chronology in this region. In recent years, detailed reports have been published on the materials from the Fukuda Shellmound and the Takashima-Kurotsuchi site; these materials were collected by Sugao Yamanouchi and used as typical specimens in his Late and Final Jomon chronology (Izumi and Matsui 1989, Yano ed. 2004).
The main purpose of this report is to throw light on the substance used in the Hikosaki-types of pottery in the original form as established by Yamanouchi. Fortunately, the information about the trench and layer of most of the excavated materials could be recovered by their labels and markings. As Ikehasu (1971) pointed out, sherds of different types were distinctively found from different trenches or layers at the Hikosaki Shellmound; it is quite possible that Yamanouchi emphasized on such an archaeological context in establishing his concepts regarding the Hikosaki types. Therefore, the typological classification in this report was based not only on the conventional study of the Hikosaki types of pottery but also on this archaeological context. The figures of the pottery are essentially displayed in this typological order. As a result of refitting, some of the sherds with no markings and labels were combined with those with such information. Some of the sherds from the same ware had been excavated from multiple trenches or layers; in such cases, it was assumed that the materials belonged to the trench and layer which yielded the most number of sherds. Table 5 presents detailed information on the pottery.
There exist 15723 sherds belonging to the category of Jomon pottery, weighing 120 kg in total; 9557 sherds (62 kg) belong to the Early Jomon period; 2015 sherds (22 kg) belong to the Middle Jomon period; 4147 sherds (37 kg) belong to the Late Jomon period; and four sherds belong to the Final Jomon period. From the typological point of view, these sherds were classified into 16groups. Groups 1-4 belong to the Early Jomon; Groups 5-11 belong to the Middle Jomon;Groups 12-15 belong to the Late Jomon; and Group 16 belongs to the Final Jomon period (Table 4). Groups 2 and Group 3 respectively correspond with the Hikosaki Z1 and Z2 type as mentioned in Yamanouchi’s chronology, and Groups 12 and 13 correspond with the Hikosaki K1 and K2 types, respectively. The pottery belonging to these four groups will be studied in detail later. In the following paragraphs, we will describe and illustrate briefly these 16 groups of pottery.First, we will describe the sherds unearthed from Trenches 1-13, which comprise the main body of the collection. Then, we will illustrate the sherds stored with the skeletons, the sherds excavated from Trench 14, and the sherds that are not accompanied by any information on their excavated position.
Groups 1-4 belong to the Early Jomon period. While specimens from Group 3 are the mostpredominant followed by those from Group 2 comes next, Group 4 consists of only one specimen.The descriptions of these groups will be distributed over the following six parts: (1) Trenches 11-13, (2) Trenches 5 and 10, (3) Trenches 1-3, (4) Trench 4, (5) Trenches 7 and 9, and (6) Trenches 6 and 8. The layers are largely divided into the upper and lower layers in (4) Trench 4, (5) Trenches 7 and 9, and (6) Trenches 6 and 8. In Trench 4, while Layer 3 is referred to as the lower layer, Layers 1 and 2 are referred to as the upper layers; in Trenches 7 and 9, while Layers 4 and 5 are referred to as the lower layers, Layers 2 and 3 and a pit in the Trench 9 are referred to as the upper layers; in Trenches 6 and 8, while Layer 5 is referred to as the lower layer, Layers 2, 3, and 4 are referred to as the upper layers. However, since there were some refitted examplesbetween the upper and lower layers, such upper/lower division of layers could not always be definitive.
Group 1 is characterized by thin walls and dense cord markings. This group corresponds with the Isonomori type of pottery; however, we could not find the ‘C’-shaped impressions which are characteristic of the Isonomori type. This is probably because of the small sample size. Sherds of this group were mainly found from Layer 6 of Trench 12. They can be divided into two subgroups depending upon whether they are decorated with cord markings: Subgroup A with cord markings and Subgroup B with a plain texture.
(1) Trenches 11-13 (Fig. 4: 1-11, Pl. 1a: 1-6)
Subgroup A (1-4), Subgroup B (5-6), body sherds (7-11)
1: A typical specimen of Subgroup A, decorated with ujo-jomon, or the wing-shaped cord markings on the exterior.
(4) Trench 4 (Fig. 21: 308)
308: A small body sherd with cord markings typical of Group 1.
(5) Trenches 7 and 9 (Fig. 33: 441)
441: A body sherd, cord-marked outside and scratched inside.
(6) Trenches 6 and 8 (Fig. 55: 701)
701: A body sherd, cord marked outside and smoothed inside.
Group 2 corresponds to the Hikosaki Z1 type, characterized by extremely thin walls and pressed and dragged impressions. The specimens from this group were rarely cord-marked. Because their walls were thin, impressions on the exterior formed bulges inside. There are prominent fingertip impressions on the surface. While some of the basal sherds indicate the existence of round-bottomed vessels, there is no evidence of the existence of flat-bottomed vessels. In addition to deep jars, there also exist some shallow bowls and cups. Some of the small vessels are
coloured. Sherds belonging to Group 2 were found from all the trenches except Trenches 3, 5, and 10, while they were concentrated in Trenches 11, 12, and 13. Group 2 can be classified into three subgroups according to the features of their decorations: Subgroup A, decorated with linear patterns of pressed and dragged impressions, Subgroup B, decorated with simple and rough impressions, and Subgroup C, plain, often notched on the rim.
(1) Trenches 11-13 (Fig. 5-8, Pl.1b: 12-94)
Subgroup A (12-46)
12: A vessel with three bands of pressed and dragged impressions, smoothed inside.
27: A rim sherd with cord markings and impressions, notched on the rim.
29: A rim sherd with fingernail-shaped impressions on the exterior, which is similar to Subgroup
B of Group 3. This is an unusual example from Group 2.
30 and 31: These two vessels largely resemble each other. They are both characterized by thick
walls, which is unusual for Group 2. They are also marked by a reddish body.
Subgroup B (47-53)
47: A shallow bowl with small impressions on both sides below the rim.
Subgroup C (54-85)
54: A typical specimen of Subgroup C. The rim is thicker than the body. Fingertip impressions
are clearly visible.
Body sherds (86-92, 94)
86: Body sherds, presumably near the bottom.
94: Body sherds, cord-marked outside.
Basal sherds (93)
93: Basal sherds with a round bottom, probably belonging to Group 2.
(3) Trenches 1-3 (Fig. 18: 230-233)
Subgroup A (230), Subgroup C (231-233)(4) Trench 4 (Fig. 21, Pl. 4b: 296-307)
Subgroup A (297-301), Subgroup B (302-304), Subgroup C (305-307), basal sherds (296)
296: Body and basal sherds from Group 2 with a round bottom. Most of the sherds are from
Layer 4, the lowest layer.
(5) Trenches 7 and 9 (Fig. 33, 34, Pl. 3b, 4a, 18b: 442-486)
Lower layer (442-459)
Subgroup A (442-448), Subgroup B (449-454), Subgroup C (455-457)
442: A rim sherd with a relatively thicker wall and flatter rim. ‘C’-shaped impression outside and
smoothed inside.
443, 444: Two rim sherds, cord-marked outside.
454: A rim sherd with a narrow clay band stuck outside the rim. Cord markings outside and
rough impressions inside.
Upper layer (460-486)
Subgroup A (460-471), Subgroup B (472-477), Subgroup C (481-486)
480: A rim sherd with cord markings on either side, a boat-shaped impression on the rim, and
small impressions inside.
(6) Trenches 6 and 8 (Fig. 55: 702-704)
Subgroup A (702-704)
Group 3 corresponds to the Hikosaki Z2 type, characterized by cord markings, thin walls, and tokushu-tottai-mon, or the special clay band decoration; this kind of decoration is visible in the form of narrow bands of clay pressed with an implement such as a tube-cut or split lengthwise. As a result, the pressed bands look like tubercles. Fingertip impressions are often observed on either side of the wall. There exist a small number of ceramics which are coloured red. While hourglass-shaped deep jars are predominant, there also exist some small and shallow vessels,such as bowls and cups. Most of the vessels have flat bottoms with large diameters; however,there also exist some round-bottomed specimens.
Since the specimens from this group resemble those from the Satogi I type which was established by Sugao Yamanouchi and detailed in the report on the Satogi Shellmound (Makabe and Makabe 1971), researchers have often commented that these two types are identical to each other. However, there exist some differences between these two types. In the final analysis, it can be asserted that specimens from the Satogi I type correspond with the later part of the Hikosaki Z2 type. This will be discussed in detail later.
While the sherds belonging to Group 3 are distributed all over the excavated area, they are especially concentrated in Trenches 4, 6, 7, 8, and 9. The widely varied specimens from this group will be largely classified into the following three subgroups: Subgroup A decorated with the tokushu-tottai-mon and often with one or more cord-marked flat bands around the mouth, Subgroup B decorated with fingernail-shaped impressions, and Subgroup C decorated only with cord markings (C1) or having plain surfaces (C2), often notched on the rim.
(1) Trenches 11-13 (Fig. 9, 10, Pl. 11b, 12a, 18b: 95-137)
Subgroup A (95-108), Subgroup B (109-110), Subgroup C (111-116), body sherds (117-125), basal sherds (126-137) 107: A body sherd lacking clay bands. Impressions are directly pressed and dragged on the surface with ujo-jomon or wing-shaped cord markings. 109, 110: Two rim sherds with fingernail-shaped impressions on the cord markings. The latter shows small scratchings along the fingernail-shaped impressions. 116: Sherds of a shallow bowl. Though this bowl seems to belong to the Late Jomon period atfirst glance, its rude fabric suggests that it should be included in Group 3. 121, 122: Body sherds with kessetsu-jomon or knotted cord markings. 126: A basal sherd with traces of fingernails on the exterior near the bottom.
(2) Trenches 5 and 10 (Fig. 11-17, Pl. 5, 13, 17b, 18: 138-229)
Subgroup A (138-163), Subgroup C (164-202), body sherds (203-214), basal sherds (215-229) 138: A typical specimen from Subgroup A. This is one of the largest vessels in the collection, with a quadrilateral shape and decorative motifs of narrow clay bands. The mouth is thicker than the body and the rim has a triangular cross section. While the interior surface of the mouth is smoothed, it is shaved on the lower part. 139: A small vessel with a raised bottom of a large diameter. 173: Sherds from the rim to near the bottom, covered with thick cord markings. Fingertip impressions are clearly observed. 203: Body sherds, cord-marked outside and rubbed inside. 215: Basal sherds with bivalve impressions on the edge of the bottom. Impressions of an unidentified object are observed on the bottom face. 217: Basal sherds, cord-marked outside and on the bottom face. 228, 229: Basal sherds with round bottoms.
(3) Trenches 1-3 (Fig. 18-20, Pl. 10b, 11b, 12a, 17b: 234-295)
Subgroup A (234-247), Subgroup B (248), Subgroup C (249-279), body sherds (280-282), basal
sherds (283-295) 237: A rim sherd with a small clay attachment on the rim. The attachment and the exterior are coloured red.
(4) Trench 4 (Fig. 21-32, Pl. 6, 7, 11, 12a, 14, 15, 17b, 18b: 309-440)
Lower layer (309-223)
Subgroup C (309-318), body sherds (319-323)
309, 310: Small sherds with a kind of decoration that is unusual for Group 3.
320: A body sherd with diamond-shaped cord markings.
322: A body sherd covered with knotted cord markings.
Upper layer (324-440)
Subgroup A (324-351), Subgroup B (352-355), Subgroup C (356-405), body sherds (406-416),basal sherds (417-440)
324: A typical specimen from Subgroup A, notched on the rim with fingertips. Two cord-marked clay bands are attached around the mouth. These bands were added after the surface was cord marked (Pl. 48h).
332: Sherds with prominent fingertip impressions inside.
352: A deep jar with fingernail-shaped impressions and scratchings, probably marked at the same time.
356: A deep jar with fine and orderly cord markings. The horizontal depressions inside the body show the joints of the clay bands.
370: A rim sherd cord-marked outside and on the rim, coloured red.
388, 389: Roughly made cups or bowls.
431: A convex and unstable basal sherd.
439: A basal sherd with large notches on the edge.
(5) Trenches 7 and 9 (Fig. 35-54, Pl. 8-10, 12b, 16, 17a, 18b: 487-700)
Lower layer (487-520)
Subgroup A (487-501), Subgroup C (502-512), body sherds (513-519), basal sherds (520)
487: A quadrilateral deep jar with a scalloped rim, probably with four projections in original, only two of them extant. Fingertip impressions are observed on the clay attachments on the top and bottom of the scalloped rim. The lower part of the body is reddish, affected by fire.
488: A deep jar reconstructed from small fragments. It has five flat bands around the mouth and wing-shaped cord markings.
504: A deep jar with prominent fingernail traces inside the mouth.
505: A deep jar notched on the rim with a stick-like implement. Horizontal rubbings are observed on the body surface.
Upper layer (521-700)
Subgroup A (521-565), Subgroup B (566), Subgroup C (567-622), body sherds (623-665), basal sherds (666-700)
521: A quadrilateral deep jar. An implement was pressed along the short and narrow clay bands on the rim, carving deep grooves on both sides of the bands. While fingertip impressions are observed all over the surface, fingernail traces inside the mouth and fingerprints outside the neck are especially clear (Pl. 48d, e, f, g). Most of the sherds were excavated from the upper layers of Trenches 7 and 9; however, some of the rim sherds were found in Layer 3 of Trench 4.
537: Body sherds from Subgroup A, covered with fingertip impressions inside.
566: A deep jar with two lines of fingernail-shaped impressions around the mouth. These impressions were marked from right to left.
568: Rim sherds, notched on the rim with the end of a cord (Pl. 48j).
573, 574: Deep jars covered with cord markings, leaving a plain area near the neck.
586: A deep jar, cord marked except at the plain middle part.
597: Rim sherds with prominent traces of clay bands on the exterior.
625: A deep jar, covered with fingertip impressions all over the interior. It is assumed that the fingers moved horizontally from bottom up.694: A basal sherd with bivalve impressions on the projecting edge outside the bottom.
700: Basal sherds, black, with round bottom. Fine paste.
(6) Trenches 6 and 8 (Fig. 55-62, Pl. 11b, 12a, 17b, 18b: 705-831)
Lower layer (705-714)
Subgroup C (705-710), body sherds (711-713), basal sherd (714)
705: A rim sherd with a boat-shaped impression on the rim, excavated from the bottom of Trench 6.
706: A rim sherd with obscure impressions inside.
Upper layer (715-831)
Subgroup A (715-738), Subgroup B (738), Subgroup C (740-785), body sherds (786-800), basal
sherds (801-831)
803: Basal sherds with large notches on the outside edge of the bottom, forming what is known
as the petal-shaped bottom.
Group 4 corresponds with the Tsukisaki-Kaso group. 832: Body sherds with light scratches on either side. Slant impressions are added along the horizontal groove.
The Middle Jomon pottery from the Hikosaki Shellmound is classified into Groups 5-11; most of them belong to the Funamoto type, which corresponds with the first half of the Middle Jomon period. The Middle Jomon sherds were mainly unearthed from Trenches 1-3, 5, 6, 10, and 11. They were more heavily worn than the sherds from the Early and Late Jomon periods. Materials from Groups 5 and 6 will be described as a whole, since the sherds of these groups are small in number. The descriptions of Groups 7-11 will be divided into the following three parts: (1) Trenches 10-13, (2) Trenches 1-5, and (3) Trenches 6-9.
Group 5 corresponds with the Takashima type from the early part of the Middle Jomon period, characterized by dense cord markings, ‘C’-shaped impressions on flat clay bands, and a coarse paste. Fibre impressions in the cells of the cord were unobserved in most of the specimens. 837: A body sherd with fine, short, and oblique incised lines, added between the lines of the cord markings. 839: A thick-body sherd, shaved horizontally inside. 840-841: Bottom sherds, probably quadrilateral or pentagonal in the original form.
This type corresponds to the Funamoto I A type, established on the basis of specimens from the Satogi Shellmound (Makabe and Makabe 1971). They may be distinguished from the specimens from Group 5 by their coarse cord markings, ‘C’-shaped impressions without clay bands, and thin walls. 848, 849: Decorated with zigzag impressions made by the locking method.
Group 7 consists of the Funamoto I C-I H types and the Funamoto II type (Makabe and Makabe 1971). It was difficult, however, to apply the criteria of the Satogi report to classify the Hikosaki collection. Thus, a variety of ceramics are included in Group 7, which may be largely divided into four subgroups: Subgroup A with pressed and dragged impressions on coarse cord markings; Subgroup B with clay bands, often with fingernail-shaped impressions or notches, added on coarse cord markings; the predominant Subgroup C with cord markings, often depressed or notched on the rim; and Subgroup D without cord markings, scratched or plain.
(1) Trenches 10-13 (Fig. 65-70, Pl. 19b, 20, 21: 853-919)
Subgroup A (853-872), Subgroup B (873-887), Subgroup C (888-894, 896-918), Subgroup D (895, 919)
853: Rim and body sherds with narrow bands and ‘C’-shaped impressions.
854, 855: Typical specimens from Subgroup A, decorated with pressed and dragged impressionsof ‘Σ’ shape.
888-893: Rim and body sherds with bivalve and fingertip impressions added outside the rim.
895: A rim sherd with vertical scratches outside.
(2) Trenches 1-5 (Fig. 74-76, Pl. 20b, 22: 982-1015)
Subgroup A (982-994), Subgroup B (995, 996), Subgroup C (997-1001, 1003-1015), Subgroup D (1002)
997-1002: Rim sherds with bivalve impressions on the rim.
1003: A typical specimen from Subgroup C, covered with cord markings outside and notched on the rim.
1008: A deep jar, covered with coarse cord markings outside and smoothed inside.
(3) Trenches 6-9 (Pl. 78-81, Pl. 20b, 21, 23: 1046-1084)
Subgroup A (1046-1050), Subgroup B (1051-1052), Subgroup C (1054-1082), Subgroup D (1953,1083, 1084)
1053: A deep jar reconstructed from small fragments, scratched outside and smoothed inside, with bivalve impressions on the rim.
1057: A typical specimen from Subgroup C, with a raised bottom.
1083: Rim and body sherds with scratches on either side.
1084: A rim sherd with vertical scratched lines inside the mouth.
Group 8 corresponds with the Funamoto III type (Makabe and Makabe 1971), characterized by clay bands, scratched lines, and incised lines, added on coarse cord markings. The sherds from Group 8 were relatively concentrated in Trenches 10 and 11, while they were not found in Trenches 6-9.(2) Trenches 10-13 (Figs. 70, 71, Pl. 24: 920-952)
920-927: Rim and body sherds with clay bands and scratched lines drawn with a bi-peaked implement.
929-947: Rim and body sherds with scratched lines.
948-952: Rim and body sherds with incised lines.
(2) Trenches 1-5 (Fig. 76: 1016-1018)
Small body sherds with scratched lines.
Group 9 corresponds with the Funamoto IV type (Makabe and Makabe 1971), characterized by shinsen-jomon or cord markings impressed deeply and sharrowly. While there is only one specimen (953) which definitely belongs to this group, another body sherd (1099) may also be included.
The specimens from Group 10, coarsely made in general, are characterized by scratches and rough incised lines. It contains a wide variety of ceramics which cannot be included in either the Funamoto or the Satogi II type. This group probably corresponds with the Satogi III type based on evidence from the Satogi Shellmound (Makabe and Makabe 1971).
(3) Trenches 10-13 (Fig. 72, Pl. 24b: 954-959)
(2) Trenches 1-5 (Fig. 76, Pl. 24b: 1019-1021)
(4) Trenches 6-9 (Fig. 82, Pl. 25: 1085-1093)
There is only one specimen (960) which belongs to Group 11. This is a deep jar similar to the Middle Jomon pottery found in northeast Japan. However, the paste used in it indicates that it was made locally in this region.
While these body and basal sherds cannot be clearly classified, it may be assumed that most of them belong to Groups 7 and 8. They are basically flat-bottomed, often slightly raised.
(1) Trenches 10-13 (Fig. 72, 73: 961-981)
(2) Trenches 1-5 (Fig. 76, 77: 1022-1045)
(3) Trenches 6-9 (Fig. 83, 84: 1094-1116)
Late Jomon pottery is classified into Groups 12-15. Plain sherds, which occupy 70% of the Late Jomon items, are classified as Group 15. Decorated sherds were classified into Groups 12-14; while specimens from Group 13 are the most predominant and those from Group 12 come next, only a small number of items belong to Group 14. The Late Jomon sherds were mainly unearthed from Trenches 1-5, 10, and 11, located on the central part of the terrace. The description of these groups will be divided in the following five parts: (1) Trenches 5, 10, and 11, (2) Trench 1, (3) Trenches 2 and 3, (4) Trenches 6-9, and (5) Trench 4.
Group 12 corresponds with the Hikosaki K1 type containing entai-mon pottery, or pottery with geometrical decoration on the thick rim, and surikeshi-jomon pottery, or fine pottery with erased cord markings. Most of the sherds from this group were found in Trenches 10 and 11; this is in contrast with Group 13, which will be discussed later. The sherds from Group 12 are classified into five subgroups. Subgroup A comprises entai-mon pottery characterized by incised lines in geometric motifs, scratched lines, and linear patterns made with erased cord markings. Subgroup B includes thick and large specimens with projections ornamented with an incised line and cord markings. It may be assumed that this subgroup is related to the Hirajo type, distributed around the Suo-Nada Sea and the Bungo-Suido Strait. Subgroup C includes finely-made pottery with erased cord markings edged with wide incised lines, consisting mainly of shallow vessels and bowls, sometimes coloured red. It may be assumed that this subgroup is related to the Kanezaki type found in the Kyushu region. Subgroup D includes pottery with only cord markings. Subgroup E comprises pot-shaped pottery, which may have been spouted.
(1) Trenches 10 and 11 (Fig. 85-90, Pl. 26b, 27-30, 31a: 1117-1173)
Subgroup A (1117-1146), Subgroup B (1147, 1148), Subgroup C (1149-1167), Subgroup D (1168-1171), Subgroup E (1172, 1173) 1117: A typical specimen from Subgroup A, reconstructed from several fragments. Its rim is scalloped with probably two projections. While scratched lines are observed on the exterior and on the rim, the inner surface is smooth.
1118: A deep jar with a flat rim, large in diameter. A typical specimen from Subgroup A, with zigzag motifs on the thickened rim.
1147: Scalloped rim sherds, decorated with incised lines and cord markings.
1149: A rim sherd with a bracken-shaped incised motif, also found commonly in the Kanezaki type found in the Kyushu region. The worn surface seems to have been plain without any cord markings.
1150: Red coloured sherds, polished at the area without cord markings.
1172: Roughly-made body sherds with the trace of a horizontal incised line on the upper edge.
1173: Body sherds, well polished outside and coarse inside.
(2) Trenches 1 and 5 (Fig. 98, Pl. 29a: 1300-1307)
Subgroup A (1300-1302), Subgroup C or D (1307)
(3) Trenches 2 and 3 (Fig. 108: 1466-1469)
Subgroup A (1466-1468), Subgroup D (1469)
(4) Trenches 6-9 (Fig. 111: 1525)
Subgroup A (1525)
Group 13 corresponds with the Hikosaki K2 type, characterized by thin walls, thin incised lines, and special cord markings, using knotted cords (kessetsu-jomon) or entwined cords (fuka-jo) , for instance. Sherds from this group were mainly unearthed from Trenches 1, 5, and 10.Since this group contains vessels of various forms, each item will be classified by its form at first,and then by its shape and decorations. It should be noted, however, that we often found it difficult to distinguish between small body fragments of a deep jar and a bowl, for example.
While deep jar A has a narrow mouth, a wide mon’yo-tai (decorated area), and a scalloped rim higher than that of deep jar B, the latter has a wide mouth and cord markings, often on either side. The small jar is similar to the deep jar B, differing only in its low profile and round bottom. While the shallow vessel A is bowl-shaped and is decorated on the outside, the shallow vessel B is dish-shaped and is cord-marked inside the rim. The vase is a finely-made ceramic with a globular body; some of the vases may have been spouted.
(1) Trenches 10 and 11 (Fig. 90-92, Pl. 32, 34, 35b, 36: 1174-1200)
Deep jars A (1174-1176), deep jars B (1177-1179), body sherds of deep jars or small jars (1181-1193), small jars (1194-1196), bowls (1197-1200)
1177: Rim and body sherds with cord markings on the rim and the body. Areas without cord markings are roughly polished.
1193: Body sherds, cord marked with Isetsu-jomon, or a kind of cord marking impressed with a cord with two types of cells in each string (Pl. 49s).
1194: A rim sherd with a narrow incised line on the inner surface, with small dots and special cord markings, made with knotted cords.
1196: Roughly-made rim sherds and round-bottom sherds, with cord markings on either side.
1200: Finely-made rim and body sherds; the areas without cord markings are carefully polished and smoothened.
(5) Trenches 1 and 5 (Figs. 98-101, Pl. 31b, 32b, 33, 34a, 36, 37a: 1308-1342)
Deep jars A (1308-1323), deep jars B (1324-1333), body sherds of deep jars (1334-1342),bowls (1343-1346), dishes (1347-1349), vases (spouted) (1350-1354)
1308: A typical specimen of deep jar A, with a pressed and dragged groove on the rim.
1311: One of the typical specimens of deep jar A, with knotted cord markings on the bulged mouth (Pl. 49p).
1317: A rim sherd with erased cord markings; the cord markings between three grooves were erased by polishing (Pl. 49q).
1320: A rim sherd with fukusetsu-jomon, or the bi-cells cord markings.
1323: Decorated with a kind of pseudo cord markings, impressed with a rolling spiral shell.
1327: A typical specimen of the deep jar B, probably with cord markings of the isetsu-jomon type,similar to 1193 (Pl. 49s).
1337: A body sherd with fukusetsu-jomon, similar to 1320 (Pl. 49r).
1342: A body sherd with pseudo cord markings as a result of spiral shell rolling.
1350: A finely-made black rim sherd with a hint of red on the exterior. Impressions of spiral shell rolling are added between the narrow grooves.
1354: A roughly made nozzle fragment.
(6) Trenches 2 and 3 (Fig. 108, Pl. 33, 35b, 36: 1470-1484)
Deep jars A (1470-1473), deep jars B (1474-1479), body sherds of deep jars or small jars (1480-1482), small jar (1483), bowl (1484)
(7) Trenches 6-9 (Fig. 111: 1526)
1526: A rim sherd from a shallow vessel A, decorated with delicate grooves, small dots, and cord markings.
A small number of pottery specimens belong to Type 14 which corresponds to the Takehara type in a limited sense. Type 14 is characterized by the scores and the impressions of spiral shell rolling added between the incised lines, instead of cord markings. This variety of ceramics has often been classified as the Hikosaki K2 type; at the Hikosaki Shellmound, however, its distribution was different from that of the Hikosaki K2 type.
(2) Trenches 10 and 11 (Fig. 91, Pl. 37b, 38a: 1180)
1180: A small rim sherd with a narrow incised line and scores on the inside surface.
(3) Trenches 6-9 (Fig. 111, Pl. 37b, 38: 1527-1533)
1531: Body sherds with arc-motifs outside, impressed by spiral shell rolling.
Group 15 consists of plain sherds from the Late Jomon period which were excavated from the same layers as sherds from Groups 12, 13, and 14. Although Group 15 is originally associated with Groups 12, 13 and 14, the relation between them will be discussed later. The ceramics from Group 15 are covered with various imprints, left behind by scratching, rubbing, carving, and polishing. Since most of the items exist as small fragments, they will be largely divided into deep jars and shallow vessels according to their form and subdivided according to the shape of the rim.
Deep jars
Most of the deep jars from Group 15 have wide mouths. It appears that they generally have slightly constricted necks. Their surfaces are covered with features formed by scratching, rubbing, carving, or polishing, which coexist in most cases. They are classified into four subgroups according to the shape of the rim. Though it was difficult to apply these criteria strictly, we will venture into the classification in order to show the variations in their shapes. The category deep jar A is marked by thick rims and is further divided into deep jar A1 with a rim thickening on the outside and looks like a band, and deep jar A2 with a rim thickening on either sides; deep jar B has a sharp rim, deep jar C a round rim, and deep jar D a flat rim (top row of Fig. 164).Shallow vessel Compared to the deep jars, the shallow vessels more frequently display polished surfaces and are generally more detailed in fabric. The shallow vessels consist of two basic forms, bowls and dishes. Since the former is group larger, we will classify the sherds belonging to it into four subgroups, adopting the same criteria as was used to classify the deep jars. Bowl A has a thick rim, bowl B a sharp rim, bowl C a round rim, and bowl D a flat rim (Fig. 164, lower part).
(1) Trenches 10 and 11 (Figs. 93-97: 1201-1292)
Deep jars A1 (1201-1204), A2 (1204-1205), B (1207-1222), C (1223-1230), and D (1231-1248);bowls A (1249-1252), B (1253-1268), C (1269-1282), and D (1283-1288); and dishes (1289-1292)
(2) Trenches 1 and 5 (Figs. 102-106: 1355-1453)
Deep jars A1 (1355-1357), A2 (1358), B (1359-1379), C (1380-1383), and D (1384-1407); bowlsB (1408-1418), C (1419-1422), and D (1423-1439); and dishes (1440-1453)
(3) Trenches 2 and 3 (Fig. 109, 110: 1485-1522)
Deep jars A2 (1485), B (1486-1495), C (1496-1497), and D (1498-1505); bowls B (1507-1509),C (1520-1514), and D (1515-1519); and dishes (1520-1522)
(4) Trenches 6-9 (Fig. 111: 1534-1541)
Deep jars (1534-1535), shallow vessels (1536-1541)
(5) Trench 4 (Fig. 112: 1543-1549)
Deep jars (1543-1545), shallow vessels (1546-1549)
We now describe the basal sherds from the Late Jomon period which could not be classified into pottery groups. Most of them appear to be the raised-bottom fragments of deep jars.
(1) Trenches 10 and 11 (Fig. 97: 1293-1299)
(2) Trenches 1 and 5 (Fig. 107: 1454-1465)
(3) Trenches 2 and 3 (Fig. 110: 1523, 1524)
1524: A thick basal sherd, probably from a round-bottomed vessel. While the interior is smooth,the exterior is roughly worked.
(4) Trenches 6-9 (Fig. 111: 1542)
There are some sherds belonging to the class of Final Jomon pottery which will be classified as Group 16.
1550, 1551: Rim sherds from deep jars. 1552: A rim sherd from a bowl. 1553: A body sherd, possibly belonging to Group 13.
Sherds 1554-1571 were stored with skeletal remains in the National Science Museum. While the numbers used to label the skeletons and stated below follow the classification proposed by Endo and Endo (1979), they do not completely accord with the numbers used in the report by Ikehasu (1971).
UMUT-AP-HB-130116-130118 (No. 1-3) (1554-1561)
There are 197 sherds, most of which belong to the Early Jomon period. While it may be assumed that sherds 1554-1560 belong to Group 3, it is possible that rim sherd 1555 belongs to Group 2. The basal sherd 1561 belonging to the Middle Jomon period was refitted from the fragments found in Layer 2 of Trench 5.
UMUT-AP-HB-130119 (No. 4)
There are 12 sherds, all belonging to the Early Jomon period.
UMUT-AP-HB-130120 (No. 5): 1562-1564
There are 9 sherds which belong to the Early Jomon period. Sherds 1562-1564 belong to Group 3.
UMUT-AP-HB-130121 (No. 6): 1565
There are 37 sherds belonging to the Early Jomon period, except one Late or Final Jomon specimen. The rim sherd 1565 belongs to Group 3.
UMUT-AP-HB-130123-130125(Nos. 8-10): 1566
There are 15 sherds belonging to the Early Jomon period. The rim sherd 1566 belongs to Group 2.
UMUT-AP-HB-130127 (No. 12)
There are 5 sherds belonging to the Early Jomon period.
UMUT-AP-HB-130128 (No. 13): 1567
There are 3 sherds belonging to the Early Jomon period. The rim sherd 1567 belongs to either Group 2 or 3.
UMUT-AP-HB-130129 (No. 14): 1568
There are 25 sherds belonging to the Early Jomon period. The body sherd 1568 belongs to Group 3.
UMUT-AP-HB-130130 (No. 16): 1569
There are 33 sherds belonging to the Early Jomon period except two sherds belonging to the Late or Final Jomon period.
UMUT-AP-HB-130131 (No. 17): 1570
There is one body sherd (1570) belonging to Group 3 from the Early Jomon period.UMUT-AP-HB-130132 (No. 18) There is one sherd belonging to the Early Jomon period.
UMUT-AP-HB-130133 (No. 19)
There are two sherds belonging to the Early Jomon period.
UMUT-AP-HB-130134 (No. 20)
There is one sherd belonging to the Early Jomon period.
UMUT-AP-HB-130136 (No. 22): 1571
There are 12 sherds belonging to the Early Jomon period.
In the following section, we will describe the pottery specimens that are not accompanied by any information on the excavated position, together with the ceramics excavated from Trench 14 which is different from the other trenches.
It is possible that sherds 1572-1613 were excavated from the trial trench. While rim sherds 1572-1574 are classified as Group 3, sherds 1575-1778 can only be attributed to the Middle Jomon period. The body sherd 1579 may belong to Subgroup C of Group 12. Sherds 1580-1583 may be classified as Group 13. Rim sherds 1584-1610 belong to Group 15; they consist of fragments of deep jars (1584-1600) and bowls (1600-1611). Basal sherds 1611-1613 belong to the Late Jomon period.
A small number of Jomon ceramics were excavated from Trench 14. While body sherds 1614 and 1615 are classified as Group 2 from the Early Jomon period, sherds 1616-1619 probably belong to the Late or Final Jomon period.
Sherds 1620 and 1621 were found in an unidentified pit. Fragments of the same vessels were distributed in Trenches 6 and 9.
These are sherds not accompanied by any information on the trenches and layers excavated, classified as Group 3.
There are several small plate-shaped artifacts (1626-1630), one of which is marked with cord markings common to Middle Jomon pottery (1628). Another specimen with sparse stickings with a tubular implement (1629) may be a identified as a basal sherd from the Early Jomon period.There are rough scratchings on the surface of 1630. Specimen 1631 is probably a ball-shaped artifact with a hollow, broken into half.
(Shinji Yamasaki)
Stone artifacts excavated from the Hikosaki Shellmound were classified into 18 types according to their form. Table 6 tabulates the number of specimens in each type. We drew to scale all the specimens of the major tool types and selected pieces of retouched flakes and used flakes (Table 7). (Please refer to the catalogue for details.) Since a large number of stone artifacts lacked labels or markings, we could not get any information about the trenches and layers from where they were excavated.
While most of the flake tools are made of sanukite, there also exist some rare andesite examples. We could not find any specimens made of Obsidian and chart, which are often used in the surrounding areas. According to our macroscopic observation, most of the sanukite was brought from Mt. Kanayama. Pebble tools are mainly made of andesite and granite; it may be noted that green schist was selectively used for stone axes.
There are nine arrowheads, including one small, complete specimen (2001), two broken specimens (2002-2003), and six unfinished pieces (2004-2009). The convex-base example (2003) may also have been unfinished. Retouches of the specimens 2006-2008 are added mainly on the tip. Specimen 2009 is considered to be the unfinished example of an unusual arrowhead with a thin body.
Four specimens may be classified as points. Two unfinished examples with a sharp tip leave the cortex surface (2001-2002). Two relatively well-shaped specimens may be the middle part of point (2012-2013). We can observe the traces of knapping on the fracture of the specimen. In Okayama prefecture, only a few points have been excavated belonging to the Early Jomon period onwards. While some points from the Early Jomon period were reported at the Isonomori and Hunakura Shellmounds near Hikosaki (Ikeda and Kamaki 1951, Kurashiki Archaeological Centre 1999), they were not found at the Satogi Shellmound that pertains to the Midde Jomon period; therefore, it may be assumed that the these points belong to the Early Jomon period.
We identified two drills, one shaped like a triangular prism and made on a spall (2014) and the other retouched on the side of a wide flake.
A rectangular tool made of a wide flake with the cortex surface still remaining. We will call it a ‘spatula tool’ based on its shape, though it is assumed to be a kind of scraper.
A flake notched on the side.
A stone tool with four protuberances made of a thin flake. There are no similar examples in this region. The excavated location (30 cm depth, Trench 7) indicates that this specimen is associated with the Hikosaki Z2 type of pottery from the Early Jomon period.
We identified 18 tanged knives, including broken or unfinished ones. They consist of wide-type (2019-2028), and elongated-type (2030-2032) specimens, and a racket-shaped item as well (2029). Two small tanged knives (2003-2004) were associated with the skeletal remains No. 5. There are two fragments of tangs (2035-2036) and one specimen retouched only around the tang, which is probably unfinished (2037).
It has been noticed that the structure of the tanged knife in this region changes chronologically. The wide-type tanged knife is predominant in the specimens from the Early Jomon period, the elongated type displaces it in the Middle Jomon period, and the tanged knife itself becomes rare in the Late Jomon period. The specimens from the Hikosaki Shellmound demonstrate the same trend. It may also be pointed out that the elongated type is composed of a flake that is thinner than the one noticed in the wide type. Two wide types composed of relatively thinner flake (2027 and 2028: from Layer 3 and 4 of Trench 6) probably belong to the Middle Jomon period, considering their excavated locations. These two types of tanged knives also differ in the cross section of their edges: while most of the wide-type specimens are retouched from one side, all the elongated-type specimens are retouched from both sides.
Scrapers are of various shapes, but many of them are composed of wide flake and classified under the wide type. Single-edged and double-edged specimens have been identified, but they might reflect different stages of reduction. 2037 is well-shaped in plan, and might be classified as a knife since the top of it is broken. The surfaces of 2039 and 2047 show double patina (the deep-tone area represents an older stage of reduction). The blank of 2039 might have been flaked from a reused, older core. 2047 may be described as a rejuvenated abandoned scraper.
2071-2073 are typical pieces esquillees. 2078 might be classified as an example of bipolar flaking core. 2083 and 2084 are spalls. Only these specimens are evidently classified as spalls.
2085-2102 and 2104 are retouched flakes, 2013 is a used flake. 2085 and 2086 might have been made by bipolar flaking. 2093 would be better classified as an end-scraper. 2099 has a cortical surface and its negative bulb has been removed by retouching. 2001’s blank is flaked from the side of a large flake. 2101 has a polished surface and ground ridges. 2104’s blank is a flake detached from a scraper.
2107 and 2108 are conjoining flakes of slate, from Layer 1 of Trench 10. Based on their excavated location, they may be placed in the Late Jomon period. 2109 is a flake of andesite that might have been a by-product from the process of making a pebble tool.
2110-1115 are especially large flakes. Many of them have cortical or lateral butts. They might have been carried into the site as blanks.
Only 2116 is identified as core. On the surface, percussion ripples extend from the central ridge.
Three polished axes have been identified. 2117 and 2118 are composed of green schist. 2117 is a small axe with a unifacially beveled edge, and it may be viewed as a rejuvenated fragment of a larger axe. 2118 is a broken, pestle-shaped axe that might have been reused as a hammer stone. The top of it displays pecking marks. 2119 is an axe with a bifacially polished edge composed of hard shale. After it was formed into the plectrum shape, the area around the edge was polished. Considering their excavated location and physical features, 2117 (from Layer 3 of Trench 4) and 2118 (from Layer 5 of Trench 7) may be dated back to the Early Jomon period, and 2119 (from a depth of 60 cm in Trench 11) may be dated back to the Middle Jomon period.
All the specimens are knapped along both sides of the long axis. In addition, 2126 is knapped on one corner.
Specimens 2132-2136 display discoidal shape, while 2137, 2138, and 2140 have an edge on one side. Though these kinds of specimens are classified as stone weights in some reports, they are here regarded as pebble tools. Many of these are made of harder stones such as andesite and silicified stones as opposed to the general composition of stone weights. Specimen 2137 is remade from a hollowed stone with indentation.
It is often difficult to distinguish the grind stones or hammer stones from the anvil stones found at the Hikosaki Shellmound. Here, expediently, specimens that are longer than 10 cm are classified as anvil stones, and the others are defined as grind stones/hammer stones. Specimens 2141 and 2142 have pecking marks on either end and scratches on the sides. Specimens 2150-2164 are identified as grind stones, but many of them have pecking marks on the sides.
Specimens 2165-2173 are hollowed stones and 2174-2176 are stones with shallower hollows. Many of them display pecking marks on the sides. Specimens 2172 and 2173 differ from other hollowed stones insofar as the hollows of the former are apparently the result of grinding. Specimens 2178 and 2179 are pole-shaped, and 2178 appears to be worn on a lateral edge. Specimen 2179 has hollows on either side with a faint lateral striation, and some parts of it are reddened or blackened, giving it a burned appearance.
(Shinji Yamasaki)
Artifacts made of bone, antler, tusk, and shell excavated from the Hikosaki Shellmound will be described in this chapter. Refer to Tables 9 and 10 for basic information on these artifacts.
There are five fishhooks made of antler, including two unfinished examples (3001-3002). Though the fishhooks are all broken, it is certain that they all belong to the ‘large’ (5 cm to 7 cm in length) or the ‘extremely large’ (7 cm and above) variety of Jomon fishhooks.
A very thin bone spatula. It is well finished but broken at the upper end.
An antler adze with a narrow and well-polished edge.
There are four pointed tools or spears made of bone (3008-3011) and the frontal and basal fragments of spears made of ray provisions (3012-3013).
A pendant made of a shark vertebra.
Unidentified objects made of boar’s tusk.
The following attributes are described in Table 10: preservation, shell species, number of ribs, proportion of the ribs and ditches (rib＞ditch, rib=ditch, rib＜ditch), size (including the internal and external diameters along the long and short axes, height, and width of the ring), weight, polishing, etc.
Two perforated shells with two holes, made of the flat right shells of oyster, Ostrea denselamellosa, were excavated from the Hikosaki Shellmound (3019-3020). Most of the shell artifacts are finished or unfinished fragments of shell bracelets (3021-78). While 3022 is made of the left shell of Ostrea denselamellosa, all the other specimens are made of ark shells, Scapharca subcrenata.
(Ken Takahashi)
In this chapter, we will describe the stratigraphical and horizontal distribution of all the excavated potsherds, which were typologically classified and described in Chapter 4.
Fig. 151 illustrates the stratigraphic occurrence of the pottery of the Early, Middle, and Late Jomon periods in each trench. All the potsherds with information on excavated position were counted; however, since the counting was conducted before the refitting process, a single vessel may have been counted more than once. The classification into these three categories was easy even for the smallest fragments; therefore, we can vouch for the accuracy of the classification.
We now summarize the features of the pottery from each period. Early Jomon pottery is composed of a fine paste; they also have thin walls with delicate cord markings, and the fingertip impressions often remain on the surface. Middle Jomon pottery uses a coarse paste and displays loose and rough cord markings; they are also characterized by thick walls and bright colours. It may be noted that most of the Middle Jomon sherds were heavily worn on the surface. Late Jomon pottery is characterized by surface treatment, such as scratching with spiral shells and polishing, that are not observed in sherds from the Early and Middle Jomon periods.
According to Fig. 151, while specimens of pottery from the Early and Middle Jomon periods tend to be unearthed from the lower layers, Late Jomon specimens tend to be found in the upper layers. At Trenches 4, 7, 9, 12, and 13, specimens of Early Jomon pottery are always dominant. At Trench 6, while specimens from Early Jomon pottery are dominant in the lower layers, Middle Jomon specimens are dominant in the upper layers. The situation is similar in Trench 8, but there is a chronological inversion between Layer 3 and Layer 4. As a whole, however, it may be said that the stratigraphy of the Hikosaki Shellmound shows a stable condition of deposit.
Let us now turn to the distribution of Jomon pottery in detail. Here, we will largely divide Jomon pottery into three periods: Early Jomon (Group 1, 2, and 3), Middle Jomon (Group 7, 8, and 10), and Late Jomon (Group 12, 13, 14). A large number of plain potsherds (Group 15) belonging to the Late Jomon period are not discussed here, but their distribution is similar to that of the specimens from other Late Jomon groups.
Fig. 152 shows the stratigraphic occurrence of the Jomon pottery by groups. Rim sherds from each group were counted in each layer of each trench. Since sherds of a same vessel were found from multiple trenches or layers in some cases, the sum of the number of sherds was larger than the actual number.
Let us first focus on the distribution of the Early Jomon groups. In Trench 12, Groups 1 and 2 are dominant in Layers 6 and 5 respectively. Group 2 and Group 3 demonstrate different patterns of distribution; while Group 2 is mainly distributed over Trenches 11-13 in the northern part of the excavated area, Group 3 is mainly distributed over Trenches 1-10 in the south. Group 2 and Group 3 could not be separated stratigraphically in the same trench; however, it should be noted that the ratio of the items from Group 2 is relatively higher at the bottom of Trenches 4 and 9.
Next, we turn to the distribution of Middle Jomon pottery. Group 7 is dominant in Trenches 6, 10 and 11. It may be said that the distribution of Group 7 shifted to the layers upper than the ones with specimens from Early Jomon pottery. Since Groups 8 and Group 10 are small in size, we do not have enough information to examine their distribution. However, it is noteworthy that Group 7 and Group 10 belong to Trenches 6 and 8, and items from Group 8 were absent in these trenches.
We now examine the distribution of Late Jomon pottery. While Group 12 is concentrated in Trenches 10 and 11, Group 13 is mainly distributed in Trenches 1, 3, 5, and 10. Group 14 issmaller in size, but it shows a pure concentration in Trench 6.
Finally, we collate the distribution data pertaining to each group. We find a small block of distribution around Trench 12 at the stage of the formation of Group 1, which coincides with the initial period of the formation of the Hikosaki Shellmound. At the next stage, pertaining to the formation of Group 2, the distribution can be observed continuously in Trenches 11, 12, and 13;however, it is slightly extended towards the south, that is, Trenches 4, 7, and 9. At the stage of the formation of Group 3, the centre of distribution completely shifts to Trenches 4, 5, 7, and 9.
These three Early Jomon groups are mainly distributed along the cliff of the destruction. In contrast,Groups 7, 8, and 9 belonging to the Middle Jomon period are concentrated in Trenches 6, 8, 10, and 11, which correspond to the edge of the distribution of Early Jomon pottery. Group 12 belonging to the Late Jomon period is also concentrated in Trenches 10 and 11. Group 13 is mainly distributed in Trenches 1, 3, and 5 located inside the terrace; this group appears to fill the vacuum in the Early and the Middle Jomon periods.
(Shinji Yamasaki)
Hikosaki Z1, Z2, K1, and K2 types, identified and defined by Sugao Yamanouchi, are still commonly used in the chronological study of the Setouchi region. However, since Yamanouchi did not publish any paper about these types, their details still remain obscure. Here, we are going to delve into each of the Hikosaki types, based on the materials reported in Chapter 4.8.1 Early Jomon pottery types The chronological study of the pottery from the Early to the Middle Jomon periods in the Setouchi region was presented in the report on the Satogi Shellmound by Tadahiko Makabe and Yoshiko Makabe. The pottery types belonging to the latter half of the Early Jomon period were arranged sequentially as follows: Isonomori type → Hikosaki Z1 type → Satogi I type (Hikosaki Z 2 type). Following the Satogi I type, they placed a group similar to the Otoshiyama type in the Kinki region, which is regarded as belonging to the end of the Early Jomon or the beginning of the Middle Jomon period (Makabe and Makabe 1971). The Hikosaki Z1 type corresponds to Group 2 in this report and the Hikosaki Z2 (Satogi I) type corresponds to Group 3. Materials reported in this study may supplement the Satogi chronology that includes new materials. In this section, we will study the Hikosaki Z1 and Z2 types vis-a-vis the Satogi chronology.
Since specimens belonging to the Hikosaki Z1 type have been found only in the shellmounds in the south of Okayama Prefecture, it is assumed to be a local pottery type with a limited distribution. The features of the Hikosaki Z1 type at the Satogi Shellmound (Makabe and Makabe 1971) and the Hashima Shellmound (Fujita et al. 1975) are summarized as follows.
1. Thin wall, approximately 3 mm in thickness.
2. Pressed and dragged impressions on a plain body, made with a slightly curved implement.
3. Deep impressions, forming bulges inside.
4. Clear fingertip impressions on the surface.
5. Often notched on the rim.
6. Cord markings are rarely observed.
7. There are some round-bottomed specimens.
8. There are some small, shallow vessels.
Group 2 belonging to the Hikosaki Shellmound also displays these features. In Chapter 4, we classified Group 2 into three subgroups; let us now turn to the details of each subgroup. Subgroup A, decorated with linear patterns of pressed and dragged impressions, is typical of the Hikosaki Z1 type. Scalloped rims are characteristic of this subgroup. Red-coloured pottery, which existed in the Isonomori type as well, was also found in this subgroup (Fig. 5-30, 31).
Subgroup B, decorated only with simple and rough impressions, consists mainly of small vessels (e.g. Fig. 153-47, 451). Specimen Fig. 34-480 was classified under Subgroup B; it has cord markings and boat-shaped impressions on the rim. Since this type of pottery was found from the layers lower than the location of Group 3, it may be said that they are chronologically closer to Group 2. Their distribution is, however, different from that of Group 2, since only one sherd belonging to this group was found from Trenches 11, 12, and 13 where Group 2 was concentrated. In this report, most of them were classified under Group 3; however, since this group may belong to the transition period between the Hikosaki Z1 and Z2 types, we need to examine it further in the future (Pl. 18b)
Items in Subgroup C resemble the plain pottery of Group 3; however, there are a few examples with bulging mouths, which is a characteristic of Group 3. They are often notched on the rim, a feature rarely seen in Subgroups A and B.
Next, let us discuss a few issues concerning the relation of the specimens with the surrounding regions. It has been pointed that the Hikosaki Z1 type was closely related to the final stage of the Sobata type in the Kyushu region and the Tsukisaki-Kaso group in the western Setouchi region (Maekawa 1972, Fujita et al. 1975). This view seems to be valid, because these types are similar in their decoration that is devoid of cord markings and the shape of the deep jars with round bottoms. It may be noted that specimen 832 appears to belong to what is known as Tsukisaki-Kaso group, excavated from the layer under the Tsunoi ash-fall deposit near Mt. Sanbe at the Shimoyama site, Shimane Prefecture (Yagiura 2001, Shimane Prefecture Archaeological Centre 2002). Since one of the fragments of specimen 832 was found from Layer 2 of Trench 4, it seems that this vessel was associated with the Hikosaki Z1 type or the older part of the Hikosaki Z2 type. (We will later discuss if this association was with Phase I or II.)
The Kitashirakawa-Kaso II type was distributed in this period in the Kinki region. The Kitashirakawa-Kaso II type is subdivided into IIa, IIb, and IIc, based on the stratigraphical excavation at the Torihama Shellmound; it is beleived that the Hikosaki Z1 type existed in parallel with the Kitashirakawa-kaso IIc type (Amitani 1982). Since the Kitashirakawa-Kaso IIa/IIb types and the Isonomori type are almost identical, we may assume that similar pottery types were in vogue in the Kinki and Setouchi regions in this period. Regional differences between the different styles of pottery became clearer at the next stage; the Kitashirakawa-Kaso IIc type, characterized by flat bottoms, clay bands along the rim, and cord-mark decorations, is quite different from the Hikosaki Z1 type. However, there also exist some specimens from the Kitashirakawa-Kaso IIc type with pressed and dragged impressions similar to those seen in the Hikosaki Z1 type; the relation between these two types needs to be investigated in the future.
The Hikosaki Z2 type corresponds with what has been defined as Group 3 in this report. They form the greater part of the pottery samples excavated from the Hikosaki Shellmound. It should be noted that the pottery belonging to Group 3 displays different aspects in each trench andlayer. This implies the possibility of a chronological subdivision of the Hikosaki Z2 type, which will be discussed later. We will also refer to the relations between the Hikosaki Z2 type and Satogi I types, the latter having been reported at the Satogi Shellmound (Makabe and Makabe 1971).
We classified Group 3 into three subgroups in Chapter 4; Fig. 154 shows their proportions. While Subgroup C is predominant, Subgroups A and B have smaller shares. Only a few specimens belonging to Subgroup B have been unearthed to date; however, an artifact which was reported as belonging to the Hikosaki Z1 type at the Ourahama site should obviously be included in this subgroup (Kagawa Pref. E. B. 1988: Fig. 16-75).
Group 3 consists mostly of deep jars, together with some shallow vessels and bowls. While they usually have flat rims, there also exist some samples with scalloped rims, projections, or notches. The rims thicken on the inside and are cord marked. The bottoms are usually made by winding a clay band around a clay disk; the edges are often notched or pressed with a shell. Some have cord markings on the bottom face. There also exist a few round-bottomed samples.
Fingertip impressions are often observed on the surface; this is a feature common to Group 2, but Group 3 is characterized by their thick walls and the nail traces observed in the finger impressions.
The decoration seen in the specimens from Group 3 is characterized by the cord markings, the special clay band decoration seen in Subgroup A, and the fingernail impressions found in Subgroup B. Specimens from Subgroup A display narrow clay bands often pinched on the surface with an implement; the width of the implement is narrower than the band. Fingernail-shaped impressions on specimens from Subgroup B are similar to those seen in Hikosaki Z1 specimens (29), implying a relation between them. The plain pottery belonging to Subgroup C should be viewed as the hallmark of the Hikosaki Z1 type as well. It should also be noted that Subgroups A and C include a few red-coloured items. In the following sections, we will describe in detail the assemblages of Group 3 in chronological order.
Lower layers of Trenches 7 and 9
The earliest assemblage of Group 3 specimens were found from the lower layers of Trenches 7 and 9. We will call this assemblage Phase I. Most of the items are small fragments; however, there are four relatively well-preserved vessels. They were obviously concentrated in the lower layers, though their fragments were also found from the upper layers. The chronological changes in Group 3 can be observed most sharply in Subgroup A with its special clay-band decoration. Specimens from Subgroup A from the lower layers of Trenches 7 and 9 have three or four flat bands around the mouth; the clay bands are decorated with feather cord-mark patterns (ujo-jomon). Small projections on the rim, common to the Hikosaki Z2 type, have triangular cross-sections and appear to have been pinched with fingers. There also exists a quadrilateral sample (487). Specimen 504 with a unique rim shape was classified as an item in Subgroup C1 because it is cord-marked along the rim. Specimen 505 is an example of plain pottery notched on the rim; it remarkably resembles the plain pottery from Phase II and cannot be typologically divided.
Trench 4 and the upper layers of Trenches 7 and 9
The assemblage from Trench 4 and the upper layers of Trenches 7 and 9 will be referred to as Phase II. Thought the transition from Phase I to Phase II was quite gradual, we can point out some differences between these two phases. For example, Phase II specimens often display deco-rative patterns with narrow clay bands. Small projections on the rims of the specimens belonging to Subgroup A bear semi-circular cross sections and are pinched with some kind of implements.A specimen of quadrilateral pottery still exists in this phase (521).
We could also trace some differences between the materials from the upper layers of Trenches 7 and 9 and Trench 4. Materials from the upper layers of Trenches 7 and 9 are often decorated with feather-patterned (ujo-jomon) or knot-rolled (kessetsu-jomon) cord markings. There also exist a large number of plain potsherds. Specimens from Subgroup A often display three bands around the mouth. Some specimens from Subgroup C1, such as specimens 573, 574 and 586, display a plain band at the middle part. Materials from Trench 4, on the other hand, are less decorated with cord markings of feather patterns or those formed by knot rolling. Subgroup A specimens often have one or two bands. For example, 324 has two bands with cord markings around the mouth which were added after the body surface was cord-marked; this method cannot be traced in the specimens from Trenches 7 and 9. These differences appear to reflect the chronological difference between these subgroups. We will label the assemblage from the upper layers of Trenches 7 and 9 as Phase IIa and the assemblage from Trench 4 as Phase IIb, since the former is closer to Phase I and the latter to Phase III which will be described below.
Trenches 5 and 10
The assemblage from Trenches 5 and 10 will be termed Phase III. The stratigraphical evidence of the distinction between Phases II and III is undeniably insufficient. This is reaffirmed by the existence of some conjoined potsherds between Trench 5 and Trench 4, and the discovery of some Phase III items in Trench 4. However, from the broader perspective of typological transformation, the Phase III assemblage has the traditional features of Phase II pottery as well as new features of the pottery from the end of the Early Jomon period.
Group 3 sherds in Trenches 5 and 10 have thick walls and surfaces smoothed by rubbing, so that the impressions of finger-squeezing have been rubbed out. There is a quadrilateral sample (138)with a horizontally-scraped interior. The cord markings on Phase III items are generally thicker and longer than those of Phase II; though they sometimes resemble the Middle Jomon cord markings, they are not as coarse. The mouth thickens and its cross section turns triangular or quadrilateral in shape.
The flat bands on the rim of Subgroup A almost disappear. There also exists one sample with a notched narrow-band projection (141). This appears to be similar to the notches on the rim of the Otoshiyama type of pottery. Correspondingly, it is likely that the shell impressions around the bottom edge (215) changed to the petal-shaped bottom of the Otoshiyama type, resembling the basal sherd from Trench 8 (803).
As stated above, the pottery specimens belonging to Group 3 (Hikosaki Z2 type) from the Hikosaki Shellmound can be divided into three phases: Phase I (lower layers of Trenches 7 and 9), Phase II (Trench 4 and the upper layers of Trenches 7 and 9), and Phase III (Trenches 5 and 10). It appears that the distribution of the artifacts shifted from south to north. It should also be noted that this collection does not contain any material from the final stage of the Early Jomon period which follows the Hikosaki Z2 type.
First, we will quantitatively describe the variations in the cord markings of Group 3 specimens.Fig. 156 presents the ratios of items with cord markings and their variations in each trench. ‘RL’refers to impressions by a cord twisted clockwise, and ‘LR’, counter-clockwise. The ratios of cord -marked and plain potsherds are almost equal in Trenches 7 and 9 (Phases I and IIa). Feather cord-mark patterns (ujo-jomon; RL+LR) are more commonly found in Trenches 7 and 9 as compared to Trenches 4, 5, and 10 (Phases IIb and III). In Trenches 4, 5, and 10, the ratio of plain pottery gets reduced and RL cord markings slightly outnumber their LR counterparts.
Second, we will examine the chronological relation between the Hikosaki Z2 and Satogi I types (Fig. 157: 1-7). The specimens from the Satogi I type, unearthed at the Satogi Shellmound, display one or two flat bands around the mouth and also have thick walls; sherds with feather cord- mark patterns and plain sherds are rare. Since these features correspond with the Hikosaki Z2 type of pottery belonging to Phase IIb and III, it may be asserted that the Hikosaki Z2 type precedes the Satogi I type. However, a chronological gap between the Hikosaki Z1 and Z2 types is still apparent.
Third, we will discuss the chronology following the Hikosaki Z2 type. In the Satogi chronology, a pottery group similar to the Otoshiyama type from the Kinki region was placed at the final stage of the Early Jomon period (Fig. 158). This group is different from the Satogi I type in its thick walls; smoothed interiors; thick cord markings, made mostly with RL cords; and notches on either edge of the rim. These features are more akin to those traceable in early Middle Jomon pottery.
The ‘Tai type’ (Kamaki 1950) of pottery is also similar to this group (Fig. 157: 16-19). The ‘Tai type’ is characterized by rims with triangular cross sections and delicate notches; it is different from the Otoshiyama type found in the Satogi Shellmound in their shape and the cord-mark patterns. It appears that the illustrated materials belonging to the ‘Tai type’ are closer to Phase III of the Hikosaki Z2 type. The recently reported materials from burial pit No. 1 at the Funagura Shellmound (Fig. 157: 8-15) are also similar to them (Kurashiki Archaeological Centre 1999). In conclusion, we may assume that the ‘Tai type’ is chronologically close to the Otoshiyama type, but it may be said to occupy the stage between the Hikosaki Z2 type and the Otoshiyama type. It is worth focussing on the recent excavations around Mt. Sanbe; while the Otoshiyama type and the later Hikosaki Z2 type were found from the layers above the Tsunoi ash fall deposit, the earlier Hikosaki Z2 type was found from the layers below the deposit (Kakuda 2004).
The transition from the Kitashirakawa-Kaso III type to the Otoshiyama type in the Kinki region appears to be in parallel with the transition in the Setouchi region as documented above. Miyoshi described the transformation of the Kitashirakawa-Kaso III type based on the voluminous materials from the Shitaka site (Miyoshi 1988, Kyoto Pref. Archaeological Centre 1989). Miyoshi divided the Kitashirakawa-Kaso III type into three stages, mainly based on the shape of the rims (Fig. 159 lower). These three stages established by Miyoshi roughly correspond with the three
phases of the Hikosaki Z2 type (Fig. 159 upper). There are also some regional differences between these types; for instance, while Subgroup A is predominant in the Kitashirakawa-Kaso III type, Subgroups B and C are rarely seen.
The chronology of the Late Jomon pottery in this region was also established by Sugao Yamanouchi. The chronological order shown by Yamanouchi is as follows: Fukuda K1 type → Fukuda K2 type → Hikosaki K1 type → Hikosaki K2 type → Fukuda K3 type (Kamaki and Kimura 1955). Based on Yamanouchi’s work, archaeologists have developed more detailed chronologies. The problems concerning the Hikosaki K1 and K2 types will be discussed in this section. We will alsorefer to the intermediate group between these two types. The relation between the decorated pottery and the plain pottery (Group 15) dominant in the Late Jomon period will also be discussed in this section.
The Hikosaki K1 type has been systematically studied by Chiba and Watabe (Chiba 1989, 1992, Watabe 1990). The author of this report has also considered an overview of the pottery chronology of this period (Yamasaki 2003). Group 12, corresponding with the Hikosaki K1 type that does not include plain pottery, is classified into four subgroups in this report. As described in Chapter 4, the Hikosaki K1 (Group 12) and K2 (Group 13) types show different patterns of distribution, separated north and south in Trench 10. In general, the Hikosaki K1 type is characterized by its relatively thicker wall, light colour, and RL cord markings. We cannot distinguish the basal sherds belonging to these two types; they are mostly small and raised. Items belonging to Subgroup A are identified as entai-mon pottery, or pottery with thick rims decorated with geometric motifs. They have flat or scalloped rims and the part between their necks and bodies is bent. The bodies are usually decorated with scratched lines or parallel incised
lines; there are no instances of cord markings. Subgroups B and C are considered to be related to the Kyushu region; however, it is likely that they were made locally and not brought from Kyushu. The projection on the items from Subgroup B is similar to that of the Hirajo type and Koikebaru-joso type belonging to the Kyushu region. Subgroup C is characterized by erased cord markings fringed with thick incised lines, resembling the Kanezaki type from Kyushu. The loop of 1149 is also quite similar to that of the Kanezaki type. Therefore, it may be asserted that Subgroup C and the Kanezaki type are chronologically parallel. Dish-shaped vessels belonging to Subgroup C and rarely seen in the Kanezaki type may be viewed as the product of regionalization. Subgroup D with cord markings is smaller in size. Some body fragments of this group with cord-mark patterns resemble the feather cord-mark pattern, marked by rolling the same cord vertically and horizontally (1170, 1171). This technique seems to be typical of the Hikosaki K1 type. Since Subgroup E consists entirely of small fragments, it is not possible to determine their whole shape. Recently, the chronological position of the Hikosaki K1 type has been discussed in relation to the Tsugumo A type, Yomoto type, and Tsushima-Okadai type IV from the Late Jomon period (Hirai 1993, Abe 1994). The controversy about the chronology of the Hirajo type of pottery is also related to that of the Hikosaki K1 type. Since the author has already dealt with these chronological problems before (Yamasaki 2003), we will not discuss them in detail in this paper. Pottery specimens with wide decoration bands around the rim, such as the samples from the Tsugumo A type, were found at Akutagawa, Osaka prefecture, and Kume-Kubota-Morimoto, Ehime prefecture, together with samples from the Horinouchi 1 type, which is a representative Late Jomon type of pottery belonging to northeast Japan (Hashimoto 1995) (Kurita 1989). Subgroup 2 of the Hirajo type also includes this kind of pottery (Nishida and Kamaki 1957). The Hikosaki K1 type was accompanied by the Horinouchi 2 type at the Ohama (Isawa and Morimoto 1981) and Nagai (Kagawa Pref. E. B. 1990) sites in Kagawa prefecture, and the Hiroe-hama site in Okayama prefecture (Makabe et al. 1979). Subgroup C of the Hikosaki K1 type corresponds with the Kanezakitype belonging to the Kyushu region. Thus, it seems that while the Tsugumo A type is chronologically parallel to the Horinouchi 1 and Hirajo types, Hikosaki K1 is parallel to the Horinouchi 2 and Kanezaki types.
This chronology is rendered problematic by the placement of the Hikosaki K1 type between the Tsugumo A type and Tsushima-Okadai type IV; these two types are similar to each other in the wide decoration bands around the rims with circular or wheel motifs. They are not chronologically consecutive, if the sequence of Tsugumo A type → Hikosaki K1 type → Tsushima-Okadai type IV and Yomoto type is held as being correct. However, the author believes that the genealogy of the decoration bands on the rims of the samples from the Tsuhima-Okadai group IV and Yomoto type are different from that associated with the Tsugumo A type. This issue will be discussed later, after we describe the Hikosaki K2 type.
The study conducted by Tadahiko Makabe (Makabe 1980), based on the materials excavated from the Takehara Shellmound, still occupies a fundamental position in the analysis of the Hikosaki K2 type. Here, we will examine the features of the Hikosaki K2 type in detail, comparing it with the Hikosaki K1 type, Tsushima-Okadai group IV, and the Yomoto type.
Group 13, which corresponds with the Hikosaki K2 type, generally consists of vessels of various forms, such as deep jars, bowls, shallow bowls, and vases. Since each form displays various shapes and decorations, and since there also exist various forms of plain pottery, the Hikosaki K2 type comprises a large variety of pottery forms. Compared to Hikosaki K1, the Hikosaki K2 type is thinner and more blackish in colour; its surface is smoothed by careful rubbing or polishing. It is usually decorated by fine cord markings and sharp incised lines; in some cases, small dots are added on the incised line, or the cord markings between the lines are erased. Incised lines and cord markings on the specimens from Group 13 show non-decorative motifs that follow linear and horizontal patterns. Most of them bear LR cord markings, which is obviously different from the Hikosaki K1 type with RL cord markings. The shape of the bottoms of Hikosaki K2 type pottery cannot be distinguished from that of the Hikosaki K1 type; it appears that most of them have raised bottoms, while some small jars and bowls have round bottoms. It is likely that basal sherds with coarse and small depression on the centre of the bottom face, such as 1454 and 1455, also belong to this type.
It should be noted that various type of cords were used for cord markings, which is unusual for Jomon pottery from southwest Japan. We identified specific cord markings such as kessetsu-jomon (Pl. 49p), fuka-jo (Pl. 49q), musetsu-jomon, fukusetsu-jomon (Pl. 49r), isetsu-jomon (Pl. 49s, t), and yori-modoshi; there also exist some examples with a kind of pseudo cord markings, impressed by rolling spiral shells. It is a matter of considerable curiosity that the Hikosaki K2 stage corresponds to the period immediately preceding the disappearance of cord markings from the pottery of southwestern Japan.
Next, we will describe the different forms of vessels. Deep jars belonging to the Hikosaki K2 type can be divided into two basic forms: Deep jar A and Deep jar B. While Deep jar A has a mouth curving inward with a wide decoration band of cord markings and incised lines, Deep jar B has a mouth opening straight and cord markings on the body and rim. Small jars are similar to Deep jar B, except for their thin walls and round bottoms. Shallow vessels can also be divided into two basic forms: bowl-shaped Shallow vessel A and dish-shaped Shallow vessel B. While the former is decorated outside, the latter is decorated inside. There are only a small number ofvases in this collection; 1350 is a well known specimen of the Hikosaki K2 type. This vase, which probably had a nozzle, resembles the Ichijo-Ji K type from the Kinki region (Izumi 1981).
Since only a small number of specimens belonging to the Hikosaki K2 type had been unearthed until recently, we could not fully explain all its features. However, the materials recently excavated at the Minami-Mizote site in Okayama prefecture (Okayama pref. Archaeological Centre 1996), include all the variations of the Hikosaki K2 type, which was classified in this report.
One of the important fruits of this research is the division of Group 13 (Hikosaki K2 type) and Group 14. The horizontal distribution of Group 13, often included in the Hikosaki K2 type, was different from that of Group 14, which is characterized by notches along the incised lines and arc patterns of pseudo cord markings impressed by rolling spiral shells. These features are also visible in the materials found in the Takehara Shellmound in Okayama prefecture (Chiba 1992). As Chiba points out, this type of pottery should be chronologically separated from the Hikosaki K2 type (Chiba ibid.). Since the details of the stage following the Hikosaki K2 type in this region are yet unclear, we have to consign them to future studies.
In this section, we will examine the transitional period between the Hikosaki K1 and K2 types, corresponding with the Tsushima-Okadai group IV belonging to the Late Jomon period (Abe 1994) as well as the Yomoto type (Hirai 1993) (Fig. 162). As it has been pointed out, the transition from Hikosaki K1 to K2 type includes many changes in decoration and form. This change in the middle of the Late Jomon period covers an extensive area and is not limited to this region only.
Hirai (1993) arranged the chronological transition from entai-mon pottery to Hikosaki K2 type pottery as follows: Tsugumo A type and Hikosaki K1 type → Nagai V group → Tsushima-Okadai group → Yomoto type → Hikosaki K2 type. This order is mainly based on the typological changes in Deep jar B; it may be appreciated as the most systematic chronology of the time. Therefore, our description of the transformation of pottery types in the following paragraphs will also follow this chronology.
The stage immediately following the Hikosaki K1 type appears to correspond with the materials from the Hinokuchi site in Kagawa prefecture (Kagawa Pref. E. B. 1990), the Takashima-Kurotsuchi site in Okayama prefecture (Tsuboi 1956, Yano et. al. 2004), and the Nagai III and IV types established by Watabe (Watabe 1990). These materials are characterized by their slightly thick rims; however, they are not as thick as the rims of the pottery belonging to Subgroup A of the Hikosaki K1 type. This kind of pottery probably corresponds with the ‘intermediate group’ between the Hikosaki K1 and K2 types, as recognized by Yamanouchi and Takahashi (Yano et. al. 2004).
This group consists of two subgroups; one group with scratched lines on the body, close to Subgroup A of the Hikosaki K1 type, and the other group with cord markings along the rim and body, similar to Subgroups B and D. It is likely that in the later period, such differences in shape and decoration were minimised and the decorations became simpler.
Following this is the stage pertaining to the Tsushima-Okadai group IV and the Yomoto type. The origin of Deep jar B, accounting for the bulk of decorated pottery, can easily be derived from the pottery group of the previous stage. Hashimoto (1994) studied the interior of Deep jar B and concluded that its decoration was derived from the decoration around the projection seen in Subgroup B of the Hikosaki K1 type as well as the pottery from the Kyushu region. The appearanceof Deep jar A, characterized by its decorative motifs and scalloped rims, also comes into question, since its shape and decoration are quite different from those of the preceding pottery groups. It is difficult to discuss this problem based on the limited materials known to us. However, it should be noted that in the middle of the Late Jomon period, pottery types such as Kuwagai-Shimo type belonging to the Kinki region and Kitakuneyama type from the Kyushu region, including vessels similar to Deep jar A, spread over southwest Japan. The appearance of Deep jar A should be studied from a broad perspective like this. However, since Deep jar A forms only a small part of the
specimens belonging to the Yomoto type (Fig. 163), we should not read too much into the appearance of Deep jar A in this region. The presence of Deep jar A increases in the succeeding Hikosaki K2 stage.
In Chapter 4, the plain pottery from the Late Jomon period was classified as Group 15 since it was difficult to determine their chronological position. Here, we will examine the relation between the plain pottery and the decorated pottery, especially Group 12 (Hikosaki K1) and Group 13 (Hikosaki K2), based on the classification and distribution of the plain pottery.
Group 15 was divided into two forms, deep jars and shallow vessels, which were classified into several subgroups (Fig. 164). Fig. 165 shows the share of each subgroup in Trenches 10 and 11, Trenches 1 and 5, and Trenches 2 and 3. Vis-a-vis the deep jars, Subgroups B and D account for more than 80% of the specimens from Trenches 1 and 5 and Trenches 2 and 3, which is slightly higher than their share in Trenches 10 and 11. With regard to the shallow vessels, there is a clear difference in their distribution in these trenches; while the share of the dishes and Subgroups
B and C of the bowls reach 20-30% in Trenches 1 and 5 and Trenches 2 and 3, the share of Subgroups A, B, and C is higher in Trenches 10 and 11.
As discussed in the previous sections, while Group 12 (Hikosaki K1 type) is concentrated in Trenches 10 and 11, Group 13 (Hikosaki K2) is distributed over Trenches 1, 2, 3, 5, and 10. Thus, it is quite probable that the subdivision of the plain pottery corresponds with the classification of the decorated pottery, thus indicating a difference between Group 12 and Group 13.
(Shinji Yamasaki)
In this section, we will examine the long-term changes in the quantitative attributes of pottery from the Early to the Late Jomon period. Since we do not have sufficient volume of materials in all the groups for this purpose, we will focus mainly on Groups 2 and 3 of the Early Jomon period, Group 7 of the Middle Jomon period, and Groups 12, 13, and 15 of the Late Jomon period. As a rule, we counted the rim sherds from each group were counted.
Fig. 166 shows the frequency distributions of the diameters of the vessels of Groups 2, 3, 7, 12, 13, and 15. Since the diameter is closely related to the pottery form, Groups 2 and 3 can be further subdivided into three forms: deep jars, jars, and bowls. For instance, 47 and 451 are bowls belonging to Group 2; 357 and 586 are jars from Group 3; and 388 and 389 are bowls from Group 3.
While Group 2 is small in number, it contains deep jars of medium size (diameter approximately 30 cm) and small size (20 cm). The bowls from Group 2 do not exceed 20 cm by a greatmargin, and this is true for Group 3 as well. The deep jars from Group 3, Subgroup C, consist of large (approximately 40 cm), mid-sized (30 cm), and small (16-18 cm) items. There are not enough items from Subgroups A and B, but they can be said to contain at least two groups of different sizes, corresponding with the large and mid-sized items from Subgroup C. It should be noted that Subgroup A also contains some small vessels, such as 139, though they do not appear in Fig. 166. Since Group 7 (Funamoto I and II) of the Middle Jomon period consists only of deep jars, we did not subdivide this group in Fig. 166. They are sized at about 28-38 cm; thus, this group did not contain small items.
We do not have enough Late Jomon items for extensive analysis. The deep jars of Group 15 show a continuous distribution at the size range of 20-40 cm; the bowls, on the other hand, can be divided into three groups: large (over 30 cm), mid-sized (approximately 26 cm), and small (under 20 cm). This kind of large bowls is characteristic of the Late Jomon period. The data pertaining to Group 13 show that vessels of different forms, namely, deep jars, small jars, and bowls, also differ in size. Subgroup B of Group 12 is much larger than Subgroup A.
The basal sherds, which could not be classified into these groups, were largely divided into the Early, Middle, and Late Jomon types (Fig. 167). Basal sherds from the Early Jomon period, belonging probably to Group 3, are loosely concentrated at about 8-12 cm. The Middle Jomon basal sherds tend to be smaller at about 6-10 cm. The Late Jomon basal sherds are much smaller, heavily concentrated at about 4-6 cm.
Since Jomon pottery was baked without a kiln, the surface colours are not uniform. There are also various post-manufacturing factors which affect the colouration of the pottery. Even though these problems remain unresolved, we will classify the colour of the pottery into five groups: black, blackish, brown, greyish-brown, and light brown in order to draw a rough sketch. Since the colour is often different in different parts a vessel, the colour of the outside surface of the mouth is used for the purposes of our analysis. We tried as much as possible to eliminate materials which seem to have changed as a result of use or in the post-deposit process.
Fig. 168 shows the percentage distribution of the different groups of surface colours seen in Jomon pottery. While black and blackish groups are predominant in Group 2, their predominance reduces in Group 3. This trend continues in Group 7 of the Middle Jomon period and Group 12 of the Late Jomon period. In Group 13, however, black becomes resurgent and can be seen in 40% of the surfaces. In Group 15 of the Late Jomon period as well, black and blackish colours display a high incidence.
It is also difficult to conduct a quantitative analysis of the wall thickness of the excavated pottery. We measured the thickness at the mouth of the vessel, avoiding partial thickenings.
Fig. 169 presents the range of wall thickness of Jomon pottery. Groups 1, 2, and 3 of the Early Jomon period mostly consist of items that are 4 mm or lesser in thickness; more than half of the items of Group 2 are 3 mm in thickness. Group 7 of the Middle Jomon period consists of thicker items showing a gradual peak at about 5-7 mm. Most of the items belonging to Groups 8 and 9 are 7 mm or greater in thickness, though we do not have enough specimens to make further deductions. The graph for Group 12 of the Late Jomon period shows a curve similar to that for Group 7. In Group 13, more than 70% of the deep jars are 5 mm in thickness and about 90% ofthe small jars, bowls, and vases are 4-5 mm thick. This implies that the items belonging to Group 13 are thinner and less varied than those belonging to Group 12.
In this final section, we are going to discuss the variations observed in the cord markings. Since we have already examined the cord markings seen in Group 3 specimens in Chapter 8, we will now focus on Groups 7, 12, and 13 (Fig. 170).
While most of the items from Subgroups A and B of Group 7 can be labelled RL, the ratio of RL to LR is approximately 5:4 in Group C. Therefore, we may say that Group 7 is dominated by RL marks. This tendency may be traced to the later part of Group 3.
Plain and cord-marked specimens are almost equal in number in Group 12; almost all the cords used were RL. The predominance of RL cord marking in Group 12 is probably a traditional feature of the Fukuda K2 type preceding entai-mon pottery, which has been known to use RL cord markings exclusively (Izumi and Matsui 1989). Group 13 differs vastly from Group 12 in the predominance of cord-marked pottery over plain pottery, using LR cords exclusively. This group is also characterized by the wide variety of special cord markings.
(Shinji Yamasaki)
The dates of objects made of bone, antler, tusk and shell were estimated from the pottery excavated from the same layers (c.f. Chapter 7). Most of the 35 specimens with stratigraphical data were excavated from Early Jomon and Middle Jomon layers.
Since all the excavated fishhooks lack points, we cannot use the ordinary classification of Jomon fishhooks which is based on the shape of the points and barbs. However, it may be assumed that they were single fishhooks with straight shanks and knobs. While 3004 was stored with the Human Bone No. 6 which belongs to the Late Jomon period, 3002 and 3005 were associated with Early Jomon pottery. These may be viewed as rare examples of Early Jomon single fishhooks from western Japan.
Several large fishhooks with barbs on the outside have been excavated from the Late Jomon shellmounds in Okayama prefecture. An unfinished fishhook excavated from the Tsugumo Shellmound throws light on the manufacturing process of such fishhooks (Fig. 171). The specimen is 8.5 cm in length, 5.8 cm in width, has a maximum thickness of 1.3 cm, and weighs 24.22 g. The manufacturing process entails the following steps (Fig. 172): (1) preparing a triangular antler plate, (2) perforating the centre of the plate, (3) making a U-shaped object, and (4) finishing. This fishhook is assumed to have been formed at the third step. While there are a series of studies by Kaneko (1964, 1966, 1972) on the manufacturing process of Jomon fishhooks, they do not consider the fishhooks excavated from western Japan. This Tsugumo specimen may be viewed as the starting point for the empirical study of fishhooks from western Japan.10.2 Perforated shells Two perforated shells were excavated from the Hikosaki Shellmound. They are the flat right shells of an oyster with two holes, which are usually called kaimen, or ‘shell masks’. The term ‘shell mask’ has been used for a flat shell with two or three perforations. The Hikosaki examples belong to the Early Jomon period, which transcend the conventional chronological framework of ‘shell masks’ (Fig. 173). This implies that we have to distinguish between large ‘shell masks’, which could have really been used as masks, and small perforated shells, such as the Hikosaki ones.
There are 56 shell bracelets made of ark shells, Scapharca subcrenata. Since half of the specimens are unfinished, we were able to reconstruct the manufacturing process of the ark shell bracelets (Fig. 174). The process consists of perforating the apex of the shell, widening the perforation, grinding the inner edge, adjusting the outer edge, and grinding the surface. Two patterns of inner-edge grinding were identified: pattern A using a large, flat grind stone, and pattern B using a small in-hand grind stone.
Two shell bracelets with wave-like carvings on the edges were excavated from the Hikosaki Shellmound. Similar examples are widely distributed from the Korean Peninsula to the Hokkaido, from the Early to the Late Jomon period (Fig. 175). Since they show sparse and wide distribution, it is likely that such processing occurred separately in each area where shell bracelets were heavily used.
(Ken Takahashi)
This research first helped us in describing the composition of the Hikosaki types by referring to several specimens. Though the details pertaining to each type were not so different from conventional recognitions, we were able to clarify some new aspects. For example, it was revealed that the Hikosaki Z2 type, which is the most predominant in this collection, could be divided into some phases. This makes us possible to delineate a more detailed chronology of late Early Jomon pottery.
Second, we confirmed that different types of pottery were unearthed from different layers and trenches at the Hikosaki shellmound, as Ikehasu reported previously. Noticing this change in distribution is important to understand the formation process of the Hikosaki site. Most of the burials were concentrated in places dominated by Early pottery types. This indicates that these burials were probably constructed at the Early Jomon period, as reported by Ikehasu. Since there exist some idiosyncratic styles of burials at the Hikosaki site, it is necessary to discuss their chronological placement in detail.
Third, we reported various kinds of stone artifacts and artifacts made of bone, antler, tusk, and shell. The study of a large number of Early Jomon materials, such as shell bracelets, a spatula, fish-hooks, and an antler adze, are very important to understand the Early Jomon culture in this region.
(Shinji Yamasaki, Ken Takahashi)