Original Articles
Taphonomic observation of Jomon human skeletal remains from a collective burial of the Gongenbara shell-mound, Chiba Prefecture, Japan
2023 Volume 131 Issue 2 Pages 89-105
Details
2023 Volume 131 Issue 2 Pages 89-105
To evaluate burial customs in the Jomon period of Japan, we observed the taphonomic signatures of human bones from a secondary collective burial and compared them with those from individual burials. We compared the compositions of identified bone parts, degree of weathering, and damage of the bone surfaces based on two neighboring shell-mound sites, Gongenbara and Horinouchi in Chiba Prefecture; a famous secondary collective burial had previously been discovered at the Gongenbara shell-mound. The degree of weathering differed between the collective and individual burials, and also differed depending on the horizontal and vertical location within the collective burial. These differences imply that a variable degree of weathering plausibly occurred at or after the secondary burial, but not at the primary burial. Regarding bone surface damage, specific types of damage were more frequently observed on the limb bones in the collective burial than in the individual burials, although different types of damage were observed in the skull assemblage. These findings suggest that different burial processes could cause the secondarily reburied bones to be exposed to the air for a longer duration of time or more easily be accessed by insects or other small animals, which might have produced variable taphonomic signatures.
Human burial practices in the Jomon period were diverse, in accordance with its long duration (approximately 16000–2500 BP) and its wide geographical range across the entire Japanese archipelago. Jomon burial practices have been investigated in terms of their regional and temporal variations (Watanabe, 1973; Suzuki, 2010). In the case of human-bone-bearing burials, the direction of the head, burial posture of the dead, and information obtained from the recovered bones have been discussed (Yamada, 2008). If the placement of human bones in a burial pit is in anatomical order, this is called a primary burial of the corpse. If the bones are deviated from their anatomical arrangement, often with multiple individuals buried in one place, this is termed secondary burial.
Burial practices changed and developed over time in the Jomon period. In the Kanto region, reburial and multiple burial practices already occurred during the Initial Jomon period (Suzuki, 2010). By the Middle Jomon period, Hai-oku-bo (‘abandoned house burials’), where multiple individuals were placed near the floor of a dwelling pit, increased in number (Nakamura, 2018). Different hypotheses regarding the placement of the dead bodies have been proposed; one is that the bodies were placed on the floor of the dwelling pit (Saito, 1978), but another view is that some bodies were buried in sediment deposited in the dwelling pit (Horikoshi, 1986). More recently, Takahashi (2007) postulated the suspension of the residential function of the dwelling and its reuse as a graveyard, where Jomon people recognized the structure of the dwelling pit and arranged the dead bodies along the wall or avoided the pillar holes of the abandoned house (Takahashi, 2004). In the Late Jomon period, secondary collective burials with commingled bones from multiple persons, such as those found in the Gongenbara and Kosaku shell-mounds in Chiba Prefecture, spread mainly in the Kanto region (Nishizawa, 2015; Watanabe, 2015). These collective burials increased in size and scale from 4–30 individuals in three pits in the Shimo-Ohta shell-mound to remains of more than 100 individuals discovered at the Nakazuma shell-mound. Different purposes have been suggested for the mass burials of various scales. For example, collective burials in the Shimo-Ohta shell-mound supposedly represent a simple clean-up operation of human bones disturbed by ancient river erosion. By contrast, in large-scale collective burials such as in the Nakazuma shell-mound, the role of social bonding was strengthened for the community members who might share memories of their ancestors (Yamada, 2018). In addition, most of the collective burials in this region included members of every generation of the community except infants, and were often accompanied by grave markers and a pillar for the burial (Yamada, 1995; Shitara, 2007). The increasing trend of diversity in burial styles has been assumed to be related to the cold climate in the Late to Final Jomon period; it has been hypothesized that the cold weather induced group dispersion and group-size reduction for the purpose of efficient resource utilization, which made the Jomon people unable to perform the same scale of ancestral rituals as before, although several closely related groups may have gathered together periodically to construct communal cemeteries (Shitara, 1993, 2015).
Thus, research on Jomon burial has been mainly based on archeological information. In contrast, many studies have been conducted in other parts of the world from the perspective of human bone taphonomy, whereby the condition of preserved bone surfaces is described, which may help to understand the burial processes of ancient people. Human and animal body decomposition is known both empirically and theoretically to be affected by burial rituals, microorganisms, and other factors, and the appearances of remaining bones are further changed by external invaders such as plants and animals (Binford, 1987; Micozzi, 1991; William and Marcella, 2001; Eline et al., 2017). It has also been suggested that the degree of weathering or damage on bone surfaces varies through the fossilization process, and that such effects can be used to identify primary and secondary burials and other scenarios such as dying by the roadside (Mack et al., 2015; Knusel and Robb, 2016).
Weathering of bone surfaces has been classified into levels of modification; for example, five levels have been applied to animal bones exposed to the ground (Behrensmeyer, 1978), and seven levels have been applied to human bone surfaces based on the amount of wrinkling (McKinley, 2004). For bone surface traumas, in addition to human-induced cutmarks or chopmarks, other types of damage to the surfaces of human bones have also been classified and studied, such as cracks in the long bone shafts of the extremities (Shipman et al., 1981; Blumenschine, 1995; White and Folkens, 2005). For example, the long-bone fracture patterns of remains of native Americans were categorized into fresh or perimortem fractures and dry or postmortem fractures; the former were related to dismemberment for the purpose of cannibalism (White, 1992). Recently, a wide variety of examples of human-induced trauma, as well as damage caused by animal bites or indentation by plant roots, have been summarized (Fernandez and Andrews, 2016). For non-anthropogenic damage, insects such as termites and cutworms are known to produce characteristic linear marks and holes for purposes such as nesting. The shapes of the depressions, holes, and lines differ depending on the insect species. Therefore, distinguishing such insect damage from pathological or human-induced alterations can provide clues to understanding the burial process or identifying the position of body parts in the burial (Hefti et al., 1980; Martin and West, 1995; Fejfar and Kaiser, 2005; Roberts et al., 2007; Huchet et al., 2011; Huchet, 2014; Odes et al., 2016; Parkinson, 2016).
In Japan, although many studies on damage marks on ancient human bones have been reported, comprehensive studies have rarely been performed from a taphonomic perspective. Ogata (1981) reviewed antemortem and postmortem injuries. The former included an incision of a stone arrowhead piercing the right ulna of a male recovered from the Ikawazu shell-mound, Aichi Prefecture (Suzuki, 1938), whereas the latter included possible cannibalistic marks on human bones from the Omori shell-mound (Morse, 1879) and the same types of traumas from several Jomon sites (Suzuki, 1938). The cannibalistic origin of these marks has been refuted based on numerous reburial traces on bones found in recent excavations (Harunari, 1986, 2007). Regarding the appearance of bone fractures, it has been reported that natural breaks on dried bones often generate uneven and transverse fractures that are almost perpendicular, or sometimes stair-stepped, to the long axis of the bone (Suzuki, 1935).
In the present study, we observed a group of human bones excavated from the Gongenbara shell-mound as an example of collective burial in the Kanto region during the Middle to Late Jomon period, and compared these bones with those from individual burials from Gongenbara and a neighboring site, the Horinouchi shell-mound. The human bones from the Gongenbara collective burial have already been studied from a combined archeological and anthropological perspective (Watanabe, 1991): that study reconstructed the genealogical composition among the dead based on the preserved mandibles and teeth of 18 individuals and hypothesized about the burial processes on the basis of excavation data from the site. Finally, Watanabe concluded that two genealogically different kin groups immigrated to and settled at the site, and their second generation made a collective burial to bind the two kin groups (Watanabe, 1991, 2008). Although Watanabe’s study is frequently cited, his conclusions are sometimes considered too speculative to be generally accepted by researchers (Horikoshi, 2000; Yamada, 2008). In terms of reconstruction of the burial process, for example, although Watanabe recognized variability in the preservation status of the mandibular bones, he ascribed those differences to the variable degree of body decomposition in the primary burial without mentioning possible bone alteration before and after formation of the secondary burial.
In the present study, our aim was to reconstruct the reburial process of the Gongenbara collective burial from the viewpoint of taphonomy. To achieve this goal, we first assessed taphonomic indicators for the human bones and mapped them onto the burial maps made at the time of excavation. Next, we conducted a statistical analysis of the relationship between the taphonomic indicators and the positions of the human bones to describe possible bone alteration before and after the formation of the secondary burial. The taphonomic indicators applied in the present study have the potential to verify the burial process, because our observations can be linked to the in situ positions in the burial pit, which have been previously recorded, with many human bones numbered and mapped. Furthermore, we compared the frequencies of types of bone surface damage and the degree of weathering to detect the similarities or differences between collective and individual burials, which provides an osteological basis for evaluating burial practices of the Jomon period.
The Gongenbara shell-mound, located along the shore of Tokyo Bay in Ichikawa City, Chiba Prefecture, is an archeological shell-mound dating from the late Middle to early Late Jomon period (Ichikawa Archaeological Museum, 1992) (Figure 1A). The site was first discovered by Nakao Sakazume in 1948, and excavated by the Archaeological Laboratory of Meiji University in 1967. This survey by Meiji University revealed that the shell-mound was horseshoe-shaped, measuring 100 m from east to west and 70 m from north to south. Subsequently, full-scale excavations were carried out in conjunction with the 1985 land development policy (Watanabe, 1991). Nineteen dwelling pits were found at this site, with pottery of the Kasori E-IV, Shomyoji I and II, and Horinouchi I and II types, indicating archeological attribution to the Middle to Late Jomon period (Horikoshi and Ryozuka, 2008).
Locations of the sites and collective burial pit. (A) Locations of the Gongenbara and Horinouchi sites (Ichikawa Archaeological Museum, 1992). (B) Collective burial P65 at the Gongenbara shell-mound site (Watanabe, 1991).
The individual burials used in this study were Burials 1–3, excavated from the northwestern area of the site. The individual of Burial 1 was buried with a northeastern head position, and the Burial 2 individual was buried with a western head position, supine and flexed. The head and some limb bones of the individual of Burial 3 were missing as a result of later destruction, but the body was assumed to have been buried individually based on articulation from the seventh cervical to third lumbar vertebrae (Yamaguchi, 1987). Small numbers of bone fragments and teeth were recovered from both individual Burial 4 and infant Burial 6 (Watanabe, 1991), but these were not included in the present study. In the northwestern part of the site, a circular burial pit with a diameter of approximately 1.4 m, P65, was found (Yamaguchi, 1987), where a pile of human skeletal remains was buried collectively (Figure 1B) associated with Late Jomon pottery. These human bones and artifacts are stored at the Ichikawa Archaeological Museum.
The Horinouchi shell-mound, located adjacent to the Gongenbara shell-mound (Figure 1A), is a type site for Horinouchi-type pottery of the Late Jomon period. Multiple excavations between 1904 and 1954 (Koganei, 1904; Suzuki et al., 1957) uncovered human remains of approximately 18 individuals, of which 10 individuals were relatively well preserved. These specimens, which are now stored at the University Museum, The University of Tokyo, were recorded as individual burials and thus used in this study (Mizushima et al., 2006). In addition to Horinouchi-type pottery, Shomyoji, Kasori B, and Angyoh-type pottery, indicative of the Late to Latest Jomon period, were also discovered (Ichikawa City, 1987; Horikoshi and Ryozuka, 2008).
To allow comparison of the taphonomic observations between the collective and individual burials in this study, three individuals from the Gongenbara shell-mound and nine individuals from the Horinouchi shell-mound were compiled; we included the latter because having only three Gongenbara individuals is too small a number for comparison.
Minimum number of individuals at Gongenbara collective burialFirst, all the human bones were identified and classified into the following skeletal parts: skull, clavicle, scapula, vertebra, sternum, rib, humerus, radius, ulna, carpal bones, hand bones (metacarpals and phalanges), patella, femur, tibia, fibula, tarsal bones, and foot bones (metacarpals and phalanges). The side of the body from which the bones came was identified if possible. Next, referring to Krogman (1962), specimens were divided into two age categories, immature and older (including adult), based on the presence or absence of epiphyseal fusion and the size of each bone. For the major bilateral bones, we identified right-left pairs of the same individual. The minimum number of individuals (MNI) for the skeletal parts was determined.
In situ position of human bones in burial pit P65We analyzed the in situ positions of bilateral bone pairs of the same individual in the excavation record. We tested the association between the distance of the bilateral bone pairs and the degree of weathering, to infer the burial processes for the paired bones. The burial map was recorded during the 1985 excavation as five sheets of accumulated layers with a depth of approximately 10–15 cm in total, in which layer 1 was the most superficial and layer 5 was at the bottom. These sheets were digitally traced to make a superimposed map, where the placement of the left and right pairs was confirmed. The pair placements were classified as ‘close pairs’ and ‘distant pairs’ according to their horizontal proximity on the map by thresholding actual distance at 7.5 cm, approximately half of the preserved shaft length of the bilateral bones.
Degree of weathering on the bone surfaceThe degree of weathering on the bone surface has been considered to be a measure of the amount of time the bone has been exposed to the ground (e.g. Behrensmeyer, 1978). In this paper, weathering refers to progressive changes of the bone surface caused by chemical and mechanical effects. The bone surface changes gradually as a result of exposure on the ground surface and/or the soil conditions. We selected long limb bones with more than half of their surface remaining and cranial bones with approximately the same degree of preservation as the limbs. We then observed the bone surface and classified the weathering degree from A to E, mainly based on the superficial wrinkles that formed in all directions on the surface as weathering progressed (Figure 2). The wrinkles are composed of clusters of shallow grooves that develop from a zigzag or branched pattern to a meshwork or condensed pattern. The wrinkling seems to be caused by root etching and some sort of trampling, and in extreme cases, is also affected by erosion (Fernandez and Andrews, 2016). The degrees of weathering are as follows.
A no wrinkles
B multiple long wrinkles
C long wrinkles throughout the bone surface (sometimes connecting with each other to form a mesh-like pattern)
D fine wrinkles over the entire surface of the bone (forming larger condensed mesh patches), whitish color
E fine wrinkles in all directions over the entire surface of the bone, and white color
Evaluation of the degree of weathering (grades A–E). As weathering progressed (approaching grade E), numerous wrinkles ran in all directions across the entire surface, and the color of the bone surface whitened.
We also mapped the degree of weathering for the major limb bones and skulls in the burial record. To relate the arrangement of the bones with the degree of weathering, we measured the distance from each point of the skull and limb bones to the center of the pit, which was defined after the collective bones were encircled. The distances were calculated for the 15 skulls that were relatively well preserved, and for each limb bone with approximately more than half of the bone surface intact. The measured distances were tested by one-way analysis of variance and compared pairwise by multiple comparison tests. To investigate the stratigraphic effect on the degree of weathering, the frequencies tabulated with the stratigraphy were tested by χ2-tests and Spearman’s rank correlation analysis.
Damage to the bone surfaceFollowing White’s (1992) classification, we checked cutmarks, chopmarks, peeling, and carnivore and rodent bite marks. In addition, we classified other types of bone damage into eight categories as follows (Fernandez and Andrews, 2016) (Figure 3).
Categories of bone surface modification: (a) linear (single): faint, single mark (c. 5 mm in length); (b) linear (double): double (parallel, or crossed) lines; (c) linear (V): longer (over 1 cm), V-shaped in cross-section; (d) linear (U): longer and broader, U-shaped in cross-section; (e) pit (smooth): oval-shaped, smooth-floored pit; (f) pit (rough): oval-shaped, roughly jagged-floored pit; (g) pit (irregular): irregularly shaped, roughly jagged-floored pit; (h) pit (hole): small, round hole penetrating the cortical bone.
The first categories consist of four types of linear marks, (a)–(d). Damage (a) is a linear (single), fine scratch approximately 5 mm long. This mark was most abundant. Ten or more marks were found in many limb bones, and they were scattered over the entire surface of the bone. Type (a) was often absent or present in small numbers in smaller bones such as the patella or foot and hand bones. Damage (b) consists of double linear scratches running close together in parallel, or two lines intersecting at a point, forming a V shape. This type was found mainly with type (a), but never in numbers of more than 20 and usually fewer than five. Type (b) was often absent on the scapula, as well as smaller bones such as the patella or hand and foot bones. Damage (c) is linear damage, approximately more than 1 cm in length with a V-shaped cross-section. This type was found in small numbers (usually one to three) in the upper limb bones, but never in the lower extremities. Damage (d) is similar to but broader than type (c), and has a U-shaped cross-section. This damage type was seen in small numbers in some limb bones, and was not found in hand or foot bones.
Depressions that differ in shape from carnivore or rodent bite marks as described by White (1992) were classified as four types of pits, (e)–(h); these damage types may have been caused by insects or other small animals. Similar pits or holes have been assumed to have been made by termites or cutworms who passed through the bones to build their nests or gnawed on the bones (Britt et al., 2008; Fernandez and Andrews, 2016). Pit (e) is an oval-shaped depression without jagged striae at the bottom or around the depression. This pit type was often found in clusters, with the number of pits varying from one to five or more within each cluster in some limb bones. Type (e) was also not found in hand or foot bones. Pit (f) is an oval-shaped depression with jagged striae. These pits were found only on the extremity bones, and in numbers fewer than five. Pit (g) is a non-circular (e.g. polygonal or stellate) depression with a jagged and uneven bottom. One to three pits were found mainly on the skull and limb bones. In a small number of skulls and limb bones, five or more (g)-type pits were found in a cluster, not scattered over the entire surface. Pit (h) is a circular (1–2 mm diameter) hole penetrating the dense bone, likely marked by insects. The number of this type varied, from one or two in many bones to more than five in some. None were found in the scapula, pelvis, or hand and foot bones.
Because the numerous small bone fragments inflate the count, we limited observation to the limb bones with more than 50% of the bone surface preserved. For the skulls, most of which were fragmented, we used a combined segment of the skull; that is, the frontal, parietal, and occipital bones as a single unit.
As with the study of the degree of weathering, we examined the distribution of the damage within the collective burial pit. Using the actual distances from the center of the burial, we tested the mean distance between the presence and the absence of each type of damage by t-tests. In addition, we tested the differences in the frequencies of the types of damage between strata by χ2-tests.
Limb shaft breakage patternWe selected six long bones (humerus, radius, ulna, femur, tibia, and fibula) and classified old fracture surfaces near the midshaft into the following seven categories (White, 1992; Shipman et al., 1981) (Figure 4):
ST (step): breakage columnar or like a step
SW (sawtooth): cross-section jagged like a sawtooth
VS (V-shaped): both ends of the diaphysis are more concave than the center, creating a V shape
FL: fracture edge is flanked with the outer surface delaminated; this fracture possibly occurs as a result of gradually strengthened bending force to the shaft
IP (irregular perpendicular): breakage roughly perpendicular to the long axis but the surface is irregularly jagged
TP (typical perpendicular): broken surface is perpendicular to the long axis of the bone and the surface is flat
TS (typical spiral): broken edge of the shaft is obliquely spiral
Limb shaft breakage pattern (reproduced from Shipman et al., 1981: Figure 2). Limb bones that were broken near the center of the long axis and had old fracture openings were selected for observation and classified into seven types.
To examine the relationship between each fracture category and the placement of the bones, we tested the mean distance from the center of the burial to the bone for each category by one-way analysis of variance. The frequency of each category was tested based on the stratification by the χ2-test.
Table 1 shows the numbers of identified human bones from burial pit P65 and those of individual burials from both the Gongenbara and Horinouchi shell-mounds. In the collective burial, 1577 out of a total of 1849 fragments were identified as skeletal body parts. The most abundant bones were skull fragments. Bones of the whole body, as well as the upper and lower limbs, were identified, including hand bones, ribs, and vertebrae, except for the sternum. For the individual burials at Gongenbara site (Nos. 1, 2, 3), although they were partially preserved, the bones of Nos. 1 and 2 included ribs and vertebrae in addition to the skull and limb bones. Those of No. 3 were mainly vertebrae and ribs. For the individual burials at the Horinouchi site, nine out of the ten individuals yielded skulls, and five of the individual burials preserved limb long bones. Ribs and vertebrae were found in five of them, and particularly in UMUT 130831 and 130836, bones of almost the entire body, including the metacarpals, metatarsals, and other hand and foot bones, were found.
| Gongenbara shell-mound | Individual burials at Horinouchi shell-mound | Total of individual burials | P-value of χ2-test | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| P65 collective burial | Individual burials | ||||||||||||
| N | % | % without skull a | N | % | % without skull | N | % | % without skull | N | % | % without skull | Without skull | |
| Skull | 615 | 39.00% | — | 171 | 30.76% | — | 213 | 33.28% | — | 384 | 32.11% | — | — |
| Clavicle | 21 | 1.33% | 2.18% | 3 | 0.54% | 0.78% | 9 | 1.41% | 2.11% | 12 | 1.00% | 1.48% | 0.27 |
| Scapula | 41 | 2.60% | 4.26% | 15 | 2.70% | 3.90% | 14 | 2.19% | 3.28% | 29 | 2.42% | 3.57% | 0.46 |
| Sternum | 0 | 0.00% | 0.00% | 0 | 0.00% | 0.00% | 7 | 1.09% | 1.64% | 7 | 0.59% | 0.86% | 0.00 |
| Rib | 103 | 6.53% | 10.71% | 112 | 20.14% | 29.09% | 109 | 17.03% | 25.53% | 221 | 18.48% | 27.22% | 0.00 |
| Vertebrae | 88 | 5.58% | 9.15% | 96 | 17.27% | 24.94% | 138 | 21.56% | 32.32% | 234 | 19.57% | 28.82% | 0.00 |
| Humerus | 60 | 3.80% | 6.24% | 9 | 1.62% | 2.34% | 6 | 0.94% | 1.41% | 15 | 1.25% | 1.85% | 0.00 |
| Radius | 52 | 3.30% | 5.41% | 6 | 1.08% | 1.56% | 10 | 1.56% | 2.34% | 16 | 1.34% | 1.97% | 0.00 |
| Ulna | 26 | 1.65% | 2.70% | 8 | 1.44% | 2.08% | 10 | 1.56% | 2.34% | 18 | 1.51% | 2.22% | 0.51 |
| Hand bones | 127 | 8.05% | 13.20% | 23 | 4.14% | 5.97% | 10 | 1.56% | 2.34% | 33 | 2.76% | 4.06% | 0.00 |
| Carpal bone | 20 | 1.27% | 2.08% | 4 | 0.72% | 1.04% | 16 | 2.50% | 3.75% | 20 | 1.67% | 2.46% | 0.58 |
| Femur | 111 | 7.04% | 11.54% | 15 | 2.70% | 3.90% | 12 | 1.88% | 2.81% | 27 | 2.26% | 3.33% | 0.00 |
| Tibia | 64 | 4.06% | 6.65% | 7 | 1.26% | 1.82% | 7 | 1.09% | 1.64% | 14 | 1.17% | 1.72% | 0.00 |
| Fibula | 75 | 4.76% | 7.80% | 7 | 1.26% | 0.52% | 11 | 1.72% | 2.58% | 18 | 1.51% | 2.22% | 0.00 |
| Patella | 27 | 1.71% | 2.81% | 2 | 0.36% | 0.52% | 1 | 0.16% | 0.23% | 3 | 0.25% | 0.37% | 0.00 |
| Pelvis | 31 | 1.97% | 3.22% | 33 | 5.94% | 8.57% | 17 | 2.66% | 3.98% | 50 | 4.18% | 6.16% | 0.00 |
| Foot bones | 54 | 3.42% | 5.61% | 21 | 3.78% | 5.45% | 33 | 5.16% | 7.73% | 54 | 4.52% | 6.65% | 0.36 |
| Tarsal bone | 62 | 3.93% | 6.44% | 24 | 4.32% | 6.23% | 17 | 2.66% | 3.98% | 41 | 3.43% | 5.05% | 0.21 |
| Total number of fragments identified | 1577 | — | — | 556 | — | — | 640 | — | — | 1196 | — | — | — |
| Total number of fragments | 1849 | — | — | 865 | — | — | 721 | — | — | 1586 | — | — | — |
a Because the number of the skull fragments in the Gongenbara site was inflated due to the heavy fragmentation, the percentage of the each indentified bone should be recalculated after removing the skull.
The frequencies of the identified human bones in each part of the body were compared between burial P65 and the individual burial cases by the χ2-test (Table 1). The results showed significant differences in the ribs and vertebrae, as well as in the sternum and pelvis, which were dominant in the percentages of the individual burials, whereas major limbs including hand bones and patellae predominated in the percentages of collective burial P65.
Minimum number of individualsWe counted the MNI for each part of the human bones from burial pit P65 (Table 2). For adults, the cranium (counted for the temporal bone) and fibula showed the largest MNI of 24. By adding the MNI of three for immature individuals, at least 27 individuals were included in Gongenbara burial pit P65. The skull was the most abundant component (24), but the femur, tibia, and fibula were equally numerous (20–24). For the upper limbs, an MNI of only seven was counted for the ulna, whereas the humerus and radius showed MNIs of 15 and 22, respectively. Compared with the relatively small MNI of carpal bones (8), the talus and calcaneus showed larger MNIs (12 and 17, respectively). Based on these results, and considering the presence of hand and foot bones (Table 1), the P65 collective burial included bones from all parts of the body except the sternum, but the recovery rates for each bone were variable, with bones from the lower extremities having higher recovery rates than those from the upper extremities. Among the bones of the upper extremity, the ulna and carpals had lower recovery rates than the humerus and radius.
| Minimum number of individuals (MNI) | Percentage of the MNI of 27 individuals | |||
|---|---|---|---|---|
| Adult | Immature | Adult + immature | ||
| Skull | 24 | 1 | 25 | 92.59% |
| Clavicle | 14 | 0 | 14 | 51.85% |
| Scapula | 6 | 0 | 6 | 22.22% |
| Humerus | 15 | 3 | 18 | 66.67% |
| Radius | 22 | 3 | 25 | 92.59% |
| Ulna | 7 | 3 | 10 | 37.04% |
| Carpal bone | 8 | 1 | 9 | 33.33% |
| Patella | 15 | 2 | 17 | 62.96% |
| Pelvis | 16 | 0 | 16 | 59.26% |
| Femur | 22 | 2 | 24 | 88.89% |
| Tibia | 20 | 0 | 20 | 74.07% |
| Fibula | 24 | 1 | 25 | 92.59% |
| Talus | 12 | 0 | 12 | 44.44% |
| Calcaneus | 17 | 0 | 17 | 62.96% |
| Maximum number of MNIs | 24 | 3 | 27 | 100.00% |
The map of the in situ positions of the bones in burial pit P65 (Figure 5) shows a vacant space in the center, around which the long limb bone shafts were arranged, and the skulls were placed outside of these long bones. Although the skulls were fragmented, drawings from the excavation in 1985 revealed 15 cohesive skulls that were probably not crushed at the time of burial. Several of the 15 identified skulls preserved frontal, parietal, and occipital bones arranged in proximity. Several long bone shafts were placed together in loose bundles not in anatomical arrangement.
Map of the arrangement of human bones at the time of burial in pit P65 at the Gongenbara shell-mound.
The placement of bilateral pairs of the same individual was classified as ‘close’ or ‘distant’ (Figure 6), and the degree of weathering on each bone surface was evaluated (Figure 7). Seven pairs of femurs, two pairs of tibiae, three pairs of fibulae, one pair of humeri, and two pairs of radii could be identified. Among the 15 observable pairs, ten pairs were placed in proximity and five pairs were distantly placed. The same degrees of weathering were recorded in five of the ten pairs placed in proximity, and in three of the five distant pairs (Table 3). There was no significant association between the proximity of the left and right pairs and the degree of weathering on the bone surfaces (Fisher’s direct probability P > 0.05); thus, we could not find an association between the distance of the bilateral pairs and the degree of weathering.
Placement of bilateral pairs of limb bones from the same individuals in collective burial P65. Only layers with the corresponding human bones are shown. Relatively close location pairs are femurs No. 1–5, tibia No. 1–2, fibula No. 1, and radius No. 1–2.
Map of the degree of weathering of human bone surfaces in collective burial P65.
| Proximity of bilateral limb pair | |||||
|---|---|---|---|---|---|
| Close | Distant | ||||
| Degree of weathring | Same | Different | Same | Different | |
| Humerus | 0 | 0 | 1 | 0 | |
| Radius | 2 | 0 | 0 | 0 | |
| Femur | 1 | 4 | 1 | 1 | |
| Tibia | 2 | 0 | 0 | 0 | |
| Fibula | 0 | 1 | 1 | 1 | |
| Limbs in total | 5 | 5 | 3 | 2 | |
The weathering status of the bones from burial pit P65 were rated as B to E; A was not found (Table 4). Both the upper and lower limb bones had the highest percentages of B ratings. The clavicles, scapulae, and ribs had the highest percentages of C ratings. The hand and foot bones had the highest percentages of E ratings, even though the carpal and tarsal bones were generally rated as B. In contrast, all the bones from the individual burials of the Gongenbara and Horinouchi shell-mounds were graded A (data not shown), indicating almost no weathering. The degree of weathering on the bone surfaces in collective burial P65 is illustrated in Figure 7.
| Degree of weathering (%) | Total number of fragments | ||||
|---|---|---|---|---|---|
| B | C | D | E | ||
| Skull | 31% | 26% | 12% | 31% | 615 |
| Clavicle | 25% | 55% | 10% | 10% | 21 |
| Scapula | 25% | 52% | 15% | 8% | 41 |
| Rib | 29% | 38% | 33% | 0% | 103 |
| Humerus | 60% | 8% | 17% | 15% | 60 |
| Radius | 44% | 38% | 5% | 13% | 52 |
| Ulna | 35% | 26% | 22% | 17% | 26 |
| Hand bones | 6% | 18% | 29% | 47% | 127 |
| Carpal bone | 85% | 15% | 0% | 0% | 20 |
| Femur | 46% | 22% | 22% | 10% | 111 |
| Tibia | 52% | 26% | 15% | 7% | 64 |
| Fibula | 55% | 35% | 6% | 4% | 75 |
| Patella | 67% | 17% | 11% | 5% | 27 |
| Pelvis | 67% | 9% | 14% | 10% | 31 |
| Foot bones | 11% | 30% | 23% | 36% | 54 |
| Tarsal bone | 64% | 27% | 9% | 0% | 62 |
One-way analysis of variance of the relationship between the placement of skull and limb bones and the degree of weathering showed a significant difference only for the limb bones, for which a multiple comparison test showed significant differences between the bones rated B and those rated C to E (Figure 8, Table 5). To evaluate the relationship between the depth of the bone location and the degree of weathering, χ2-tests were conducted on the limb bones, and the degree of weathering was found to be significantly different at different depths, with a P-value less than 0.05 (Table 5). The Spearman’s rank correlation coefficient between the layer and the weathering degree was −0.421, with a significant p-value less than 0.001, indicating that the more weathered bones were located in the upper layers (Table 6).
Map of the degree of weathering and results of multiple comparison tests for distance from the center.
| Observation | Bone location | Methods | Skull | Statistical significance | Limb | Statistical significance | ||
|---|---|---|---|---|---|---|---|---|
| df | P-value | df | P-value | |||||
| Weathering | Layer | χ2-test | Not applicable | — | — | S | 9 | 0.039 |
| Distance from center | One-way ANOVA | NS | 14 | 0.066 | Grade B | 3 | 0.000 | |
| Damage | Layer | χ2-test | All NS | 3 | 0.3–0.9 | Damage (g) | 3 | (g): 0.000 |
| Distance from center | t-test | All NS | 13 | 0.2–0.7 | Damage (b), (c) | 60 | (b): 0.035 (c): 0.008 | |
| Limb shaft breakage pattern | Layer | χ2-test | Not applicable | — | — | All NS | 3 | 0.07–0.7 |
| Distance from center | One-way ANOVA | Not applicable | — | — | NS | 23 | 0.787 | |
| Weathering grade | Layer | |||
|---|---|---|---|---|
| 1 | 2 | 3 | 4 and 5 | |
| B | 2 | 2 | 8 | 2 |
| C | 6 | 4 | 8 | 0 |
| D | 9 | 5 | 4 | 0 |
| E | 8 | 2 | 1 | 1 |
Spearman’s rank correlation coefficient = –0.421, df = 9, P = 0.000656.
We did not find any intentional cutmarks on the bones from burial pit P65. As questionable examples, linear marks with a V-shaped cross-section were found on one humerus and one ulna (Figure 9), which may be interpreted as trampling marks. The two marks are not associated with minute grooves running alongside the main cut, which have been reported as shoulder effects and considered to be characteristic of the slicing cutmarks made by lithics (Shipman and Rose, 1983). Otherwise, the two bones exhibited several faint marks along the pillar of the bone shaft. There were no chopmarks, peelings, clear carnivore or rodent marks, or burnt bones in the collective burial pit P65. We draw attention to the white coloration found in many bones in the burial pit, because burnt human bones vary in color from black at low temperature (below 700°C) to white at temperatures above 800°C (Absolnova et al., 2012). However, we did not find any of the cracking, shrinkage, or distortion in the bone morphology that is usually associated with high-temperature burning; thus, we concluded that the white coloration was not caused by burning.
Linear marks (red arrows) with V-shaped cross-sections observed on bones from burial pit P65: (A) humerus (Specimen No. 511); (B) ulna (Specimen No. 197)
Damage to the bone surface (types (a)–(h) in Figure 3) was commonly found on the skull and extremities and less often on the carpals, tarsals, and patellae. For each case, we compared the frequency of damage in collective burial P65 and that of bones from the individual burials, and performed the Fisher’s exact test (Table 7, Figure 10).
Percentages of damaged bones (types (a)–(h)) in the collective and individual burials. Comparisons were made separately for the skull (top) and limb bones (bottom).
*Significant at the 5% level based on the Fisher’s exact test.
| Skulls, damage type | P65 collective burial (N = 32) | Individual burials (N = 29) | P-value for Fisher’s exact test | |||
|---|---|---|---|---|---|---|
| No. with damage | Percentage of skulls with damage (%) | No. with damage | Percentage of skulls with damage (%) | |||
| a | 5 | 15.63% | 19 | 65.52% | 0.000 | |
| b | 5 | 15.63% | 10 | 34.48% | 0.136 | |
| c | 2 | 6.25% | 0 | 0.00% | 0.493 | |
| d | 3 | 9.38% | 0 | 0.00% | 0.239 | |
| e | 5 | 15.63% | 3 | 10.34% | 0.710 | |
| f | 0 | 0.00% | 1 | 3.45% | 0.475 | |
| g | 5 | 15.63% | 0 | 0.00% | 0.054 | |
| h | 3 | 9.38% | 0 | 0.00% | 0.239 | |
| Limb bones, damage type | P65 collective burial (N = 62) | Individual burials (N = 39) | P-value for Fisher’s exact test | |||
| No. with damage | Percentage of limb bones with damage (%) | No. with damage | Percentage of limb bones with damage (%) | |||
| a | 35 | 63.64% | 22 | 56.41% | 1.000 | |
| b | 22 | 40.00% | 6 | 15.38% | 0.039 | |
| c | 9 | 16.36% | 3 | 7.69% | 0.361 | |
| d | 21 | 38.18% | 0 | 0.00% | 0.000 | |
| e | 22 | 40.00% | 0 | 0.00% | 0.000 | |
| f | 4 | 7.27% | 3 | 7.69% | 1.000 | |
| g | 18 | 32.73% | 0 | 0.00% | 0.000 | |
| h | 8 | 14.55% | 0 | 0.00% | 0.022 | |
In the case of skulls, the frequencies of damage types (a) and (g) were different between the collective and individual burials. The frequency of linear damage type (a) was higher in the individual burials, whereas that of irregularly shaped pits (type (g)) was significantly higher in the collective burial (although the frequency was 15.6%) than in the individual burial (Table 7). In the case of limb bones, statistically significant differences were found in many of the damage categories, except for types (a), (c), and (f), and damages were found more commonly in burial pit P65 than in the individual burials (Figure 10).
To visualize the damage types in the bone locations of the collective burial pit, we overlaid the types of damage ((a)–(h)) on each layer of the bone surface weathering map (Figure 11). The t-tests of the distance between the presence or absence for each damage type showed significant differences only for damage of types (b) and (c) on limb bones (Table 5). For both of these damage types, the mean distance from the center was smaller for damaged bones than for bones with no damage. This finding indicates that the bones with damage of types (b) and (c) tended to be placed closer to the center. However, when these results are considered together with the skull data, there seem to be no remarkable trends in the bone locations with respect to the presence of damage (Table 5).
Overlay map of bone surface damage (types (a)–(h)) and the degree of weathering in pit P65. The numbers of damage occurrences are not reflected.
Furthermore, χ2-tests for association between the layers and each damage type showed significant differences, with a P-value of 0.00023 only for damage type (g) (Table 5). Although the pairwise comparisons did not indicate any significant differences, combined samples of the upper layers (1 and 2) were shown to involve statistically more damage of type (g) than those in the lower layers (3 and 4) (P = 0.00027).
Limb shaft breakage patternsTable 8 and Figure 12 show the results of the shaft breakage patterns for the long limb bones. Of the 182 observed limb bones in collective burial P65, the sawtooth (SW) pattern accounted for the largest percentage (33%), followed by irregular perpendicular (IP) and stepped (ST). The percentages of V-shaped (VS), flanked (FL), and typical perpendicular (TP) were all low, accounting for only 3–4% each. The typical spiral (TS) pattern accounted for 15%. In 21 limb bones from individual burials, the ST and IP breakage types were abundant, whereas the VS, FL, TP, and TS types were not present. In the comparison of the two groups, a significant difference was found only in the ST, which was more common in the individual burials than in burial P65.
| P65 collective burial (N = 182) | Individual burials (N = 21) | P-value for Fisher’s exact test | ||||
|---|---|---|---|---|---|---|
| Limb bones with fracture | Percentage of limb bones with fracture (%) | Limb bones with fracture | Percentage of limb bones with fracture (%) | |||
| ST | 36 | 20% | 10 | 48% | 0.010 | |
| SW | 60 | 33% | 3 | 14% | 0.087 | |
| VS | 6 | 3% | 0 | 0% | — | |
| FL | 7 | 4% | 0 | 0% | — | |
| IP | 39 | 21% | 8 | 38% | 0.102 | |
| TP | 6 | 3% | 0 | 0% | — | |
| TS | 28 | 15% | 0 | 0% | 0.086 | |
Percentages of shaft breakage patterns in the collective and individual burials.
*Significant at the 5% level based on Fisher’s exact test.
One-way analysis of variance of the bone positions (distance from the center) among each fracture type in burial P65 did not show any difference (P = 0.787), and the χ2-test for the layers also did not show any significance (Table 5).
Human bones in burial pit P65 were accumulated in a cohesive layer within a thickness of 10–15 cm (Watanabe, 1991). Among the identified bones, skull bones were most abundant. This may have been because of heavy fragmentation of the skulls. Fifteen skulls among 24 MNI adults were found as a portion of the single cranium, which was mostly composed of anatomically continuous fragments; the main parts of the skull, such as the frontal, parietal, and occipital bones, were also identified. Thus, these skulls are assumed not to have been fragmented prior to the accumulation but in subsequent secondary burial processes. Aside from the skull bones, the hand bones were the most commonly recovered in terms of the number of fragments, where the frequencies of the identified bone parts were statistically different between collective burial P65 and the individual burials, based on the χ2-test. The recovery of the hand bones may be simply because of their small size and large number, but also may have been affected by selection of bone during formation of the secondary burial: it has been suggested that large, strong long bones tend to be more commonly recovered than fragile scapulae and fine hand bones in secondary burials (White, 1992). In the context of the MNI numbers, the recovery rate varied depending on the body parts, with higher recovery rates for the lower extremities than for the upper ones. This finding means that the reburial process in the P65 pit was somewhat biased in terms of body parts, although the bones of the whole bodies may have been carefully recovered and reinterred in the secondary burial pit.
Degree of weathering in burial pit P65For the degree of weathering, all bones from the individual burials were graded A, but those in burial P65 were graded B–E. Two reasons can be assumed for this observed difference. One is an effect of the local soil environments of the burials, although chemical data were not available for comparison. The other is because of the duration of exposure of the human bones. The latter reason would have been related to human burial practices; in other words, the more weathered conditions in burial pit P65 can be ascribed to a relatively longer period of time for exposure of these bones than those in the individual burials. When we examined the relationship between the placement of human bones and the degree of weathering in the collective burial, the B-rated bones of the extremities were more centrally distributed than the C- to E-rated bones. In addition, the bones in the upper layers were more weathered than those in the lower layers. The concentration of bones in the burial pit may have enhanced the effect of calcium to keep the bones intact, increasing the percentage of less weathered bones in the central and lower positions in the burial. The observed weathering effect would thus appear after the formation of the secondary burial, although we cannot rule out some degree of weathering after the removal from the primary burial. Following this, we should note that our observation contrasts well with Watanabe’s hypothesis on the burial process (Watanabe, 1991, 2008), in which he ascribed the variable bone surface conditions to the varying degree of decomposition of the bodies at the time of the bone accumulation from the primary burials.
Damage on human bone surfacesIn several of the damage categories (types (a)–(h)), the frequencies were significantly different between burial P65 and the individual burials. In addition, the different combinations of these categories were found to show significant differences between the skull and limb bones. The latter may have occurred simply as a result of the differences in fragmentation of the bones. Because the P65 skull samples were heavily fragmented compared with those from the individual burials, we counted the damage frequencies of the frontal, parietal, and occipital bones as separate units. For the limb bones, in contrast, the bias caused by bone fragmentation was assumed to be minimal, because we limited our observations to relatively well-preserved limb bones with more than half of the bone shaft preserved.
More types of damage were found in the limb bones from P65 collective burial than those in the individual burials. Damages similar to the observed damage (types (a)–(h)) in the P65 collective burial have been considered to be caused by insects or small animals such as moles in the soil (Fernandez and Andrews, 2016). If we accept these inferences, the collective burial may have been more accessible by these creatures than the individual burials were. For the relationship between the presence of damage and bone placement in the pit, no significant patterns were found overall in the skull and limb bones, although types (b) and (c) were found medially, and damage of type (g) was found more commonly in the upper layer in collective burial P65. These types of damage might have been affected by the placement of the bones in the pit: medial damage (types (b) and (c)) may be related to a plausible pillar hole, and the damage of type (g) in the upper layer may have resulted from that area’s higher accessibility to biting creatures.
Limb shaft breakage patternsThe most frequently recorded pattern of limb shaft breakage in collective burial P65 was jagged (SW), whereas in the individual burials, it was the stair-stepped (ST) pattern. Neither type of fracture was caused by the breakage of fresh ‘green bone,’ where the muscles and tissues were not completely decomposed and the bones were not dry; instead, postmortem processes seemed to be the cause (White and Folkens, 2005). In addition, we observed spiral fractures (TS) in the collective burial but not the individual burials, although this difference was not statistically significant. (Table 8). It is unclear whether these differences can be attributed to the different burial processes, but may have been related to the more frequent and heavier weathering of the bone surfaces in collective burial P65, which implies disposal of the bones in near-open conditions for a relatively long time. Such external factors associated with the secondary burial may have affected the breakage patterns of the long bone shafts as well as the weathering of the bone surfaces.
Characteristics of Gongenbara collective burial P65A previous study recognized and described the collective burial pit P65 as an example of a unique type of secondary collective burial, banjo-shuseki-bo, or square-shaped bone-pile burial, mainly found in the Tokai region of central Japan (Watanabe, 1991). In the in situ map, however, a square-shaped arrangement cannot be seen, and in the first, second, and third layers the long bones are arranged in various directions around a central vacant space. In the center, there was a trace of a hole with a diameter of approximately 15 cm, penetrating the layer of human bones and continuing below the bottom of the pit. This hole was interpreted as a pillar hole (Watanabe, 1991). Accepting this interpretation, it is plausible that the human bones may have been accumulated and simply placed around the pillar, resulting in the polygonal-like shape of the bone assemblage.
As mentioned above, we identified different taphonomic signatures on the bones in the Gongenbara collective burial and from the individual burials. Among the plausible factors, we may infer a stronger association of the degree of weathering with the formation of the collective burial; that is, the weathering occurred during or after the secondary reburial process, not during the primary burial (Watanabe, 1991; 2008). Other factors might also be related to the observed taphonomic signatures; for example, a pillar has been assumed to be involved in formation of the collective burial based on similar evidence of a pillar hole with a circumferential ditch from several neighboring site excavations (Yamada, 1995; Shitara, 2007). If we accept the burial site as a ceremonial place, the community members might have kept an eye on the dead and protected them from organisms such as rats and moles, but not prevented fine damage to the bones by insects or termites.
The human bone assemblage of the Gongenbara collective burial P65 consisted of almost the entire skeletons of a minimum of 27 individuals, which were reburied without maintaining anatomical order. We found no relationship between the placement of the bilateral pairs of the same individuals and the degree of weathering. Weathering was heavier in the collective burial than in the individual burials, indicating different burial processes between the two burial types. Within the collective burial, the degree of weathering changed with the bone placement in the pit; the outer bones were more severely weathered, leading to the interpretation that the secondarily reburied human bones were left exposed to the air during or after the time of reburial. This conclusion contrasts with Watanabe’s hypothesis on the formation of the collective burial. Accepting the interpretation that a monumental pillar was displayed in the center of the Gongenbara burial pit P65 and that there was a substantial time of exposure to the open-air environment, the bone accumulation may have undergone various stages of weathering. Bone surface damage in pit P65 could be attributed to the same factors in the secondary burial formation process; greater damage associated with bone accumulation was likely caused by easier access by insects or other small animals. These observations have clarified the taphonomic features of secondary burial for Gongenbara burial pit P65. This study provides further information concerning human burial customs during the Jomon period. The trend of Jomon burial practices during the Middle to Late Jomon period in the Kanto region can be examined collectively by observing the taphonomic signatures on the bones as well as interpreting them along with archeological and excavation data.
We thank Masahiro Ryozuka of the Ichikawa Archaeological Museum for providing the opportunity to study the Gongenbara Jomon human skeletal remains in his care. This study was supported in part by a Grant-in-aid for scientific research from JSPS, KAKENHI, No. 16K07530. We thank Professor Reiko T. Kono and two anonymous reviewers for improving our paper, and Sara J. Mason, MSc, ELS, and Lucy Muir, PhD, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.
