Original Articles
Taphonomic analysis of ‘scratches’ on medieval human crania from the Zaimokuza site, Kamakura, Japan
2023 Volume 131 Issue 2 Pages 69-78
Details
2023 Volume 131 Issue 2 Pages 69-78
Three excavations in the 1950s of the Zaimokuza site in Kamakura, Japan, uncovered human skeletal remains, mainly crania, of a large number of individuals buried in mass graves. Materials from the first and second excavations were previously reported as being victims of Nitta Yoshisada’s siege of Kamakura (1333) in part due to the high incidence of weapon-related trauma. Recent archaeological excavations in the area near the site have shown that it lies within a large contemporary medieval graveyard and despite consistencies in burial conditions, evidence for violent trauma, while present, was found in much lower frequencies in surrounding sites. Many of the traumas reported in the original study on the Zaimokuza skeletal remains were ‘scratches,’ described as extremely short and shallow cut marks on the crania, and it had been suggested that taphonomically induced bone surface modifications may have been mistakenly recognized as human-induced traumas. This study uses scanning electron microscopy to re-examine the Zaimokuza materials for ‘scratches’ and investigates their causes. A total of 235 examples of marks matching the description of ‘scratches’ were found on 38.2% of individuals examined that varied in their morphologies and were likely caused by different agents for different reasons. The results indicate that most were groups of ‘weak and shallow’ marks that display microscopically identifiable features commonly found on marks caused by sedimentary abrasion described in recent taphonomic studies, and so were likely caused by natural processes before burial. In addition, we found marks similar to ‘scratches’ inflicted post-excavation on clay used to restore a specimen. When taking the results into consideration, some of the interments at the Zaimokuza site may represent burial customs of the medieval period such as ‘corpse abandonment,’ rather than casualties of warfare and violence.
The Zaimokuza site, located in Yuigahama in Kamakura city, was excavated three times in the 1950s by a team from the University of Tokyo, headed by Professor Suzuki Hisashi. From the first two excavations in 1953, the skeletal remains of at least 556 individuals were uncovered, along with remains of animals such as horses and dogs, buried primarily in mass graves. Suzuki reconstructed selected crania and examined these along with a number of long bones, and the results were published in the first original report in 1956 (Suzuki et al., 1956). For reasons including the unusual burial conditions, dominance of young adult males, and high incidence of human-induced weapon trauma found on the remains, the study concluded that they likely belonged to victims of Nitta Yoshisada’s siege of Kamakura in 1333.
Since the Zaimokuza site excavations were the first discovery of large numbers of individuals buried in mass graves in Japan, Suzuki’s pioneering study has been a standard for studies in medieval Kamakura historical anthropology. However, in the half-century since, archaeological excavations in the surrounding area have recovered numerous human skeletal remains from the contemporary period, and it is now known that the site lies within the area of a large medieval graveyard (Figure 1). Despite consistencies in burial conditions across the sites, instances of violence-related weapon trauma on human skeletal remains, while present, were found to be much lower in surrounding sites (Matsushita, 2002; Hirata et al., 2004; Nagaoka et al., 2009, 2010, 2013; Sakaue, 2016). These results raised questions about the original Zaimokuza report. In particular Suzuki’s description of the types of traumas found on the Zaimokuza human skeletal remains includes a category he called ‘scratches,’ described as groups of very shallow, short cuts, which do not penetrate the external laminae of the crania, caused by a sharp weapon such as a sword. In contrast with other types of traumatic injuries found including ‘gashes’ and ‘incisions,’ and ‘stabs’ and ‘blows,’ which match modern-day descriptions of sharp-force and blunt force traumas, respectively (Walker and Long, 1977; Berryman and Symes, 1998; Boylston, 2000; Berryman et al., 2012; Arbour, 2008; Boutros-Ghali, 2008; Lewis, 2008), the description of ‘scratches’ is much more ambiguous. Suzuki reported that 153 out of 280 individuals had this type of trauma (54.1%) and it is these marks that contribute to the high overall frequency of trauma reported on individuals found at the Zaimokuza site. Suzuki’s classifications of trauma have been used in the majority of studies on human skeletal remains from medieval Kamakura (Hirata et al., 2004; Nagaoka et al., 2006, 2009, 2010, 2013; Sakaue, 2016). Table 1 shows the frequencies of the three types of sharp-force traumas reported in both Zaimokuza and surrounding sites. ‘Scratches’ in particular, are found in much less abundance at other sites. It has been supposed that Suzuki’s examination of marks on the crania did not differentiate between human-induced cut marks and other confounding factors, such as marks produced from taphonomic processes (Nagaoka et al., 2009; Sakaue, 2016).
Distribution of the archaeological sites in the Yuigahama area of Kamakura. The black line indicates the area termed by archaeologists as the Yuigahama medieval graveyard (Yuigahama Chūsei Shūdan Bochi). Black dots indicate the sites utilized in this study: 1, Zaimokuza; 2, Gokurakuji; 3, Yuigahama Chūsei Shūdan Bochi (No. 372); 4, Seiyōkan; 5, Yuigahama Minami; 6, Yuigahama Chūsei Shūdan Bochi (Seika Ichiba).
| Site name | n | ‘Gashes’ | % | ‘Incisions’ | % | ‘Scratches’ | % | Total no. of individuals with sharp-force trauma | % |
|---|---|---|---|---|---|---|---|---|---|
| Zaimokuza1 | 283 | 5 | 1.80% | 28 | 9.90% | 153 | 54.10% | 186 | 65.70% |
| Gokurakuji2 | 331 | 1 | 0.30% | 4 | 1.20% | 21 | 6.30% | 24 | 7.30% |
| YCSB* (No. 372) 3 | 592 | 6 | 1.00% | 2 | 0.30% | 3 | 0.50% | 8 | 1.40% |
| Seiyokan4 | 91 | 6 | 6.60% | 1 | 1.10% | 0 | 0.00% | 6 | 6.60% |
| YCSB* (Seika Ichiba)5 | 51 | 1 | 2.00% | 0 | 0.00% | 0 | 0.00% | 1 | 2.00% |
| Yuigahama Minami (individual burials)6 | 667 | 3 | 0.40% | 1 | 0.10% | 0 | 0.00% | 4 | 0.60% |
1Suzuki et al., 1956; 2Sakaue, 2016; 3Nagaoka et al., 2009; 4Nagaoka et al., 2010; 5Nagaoka et al., 2013; 6Hirata et al., 2004.
* YCSB, Yuigahama Chūsei Shūdan Bochi.
Research in taphonomy since the 1960s has shown that cortical bone damage which mimics human-induced cut marks may occur from an array of natural physical and diagenetic processes (Brain, 1967; Behrensmeyer, 1978; Fiorillo, 1984, 1989, 1991; Andrews and Cook, 1985; Behrensmeyer et al., 1986; Olsen and Shipman, 1988; Dunbar et al., 1989; Gaudzinski-Windheuser et al., 2010; Reynard, 2014; Pineda et al., 2014; Fernández-Jalvo and Andrews, 2016; Rozada et al., 2018). Sedimentary abrasion occurring during trampling or transport, particularly in coastal sediments with coarse sand and shells, has been shown to produce modifications on bones such as marks with broad, shallow bands of parallel striations (resembling scraping) and areas crisscrossed by abundant superficial ‘incisions’ which are similar to Suzuki’s descriptions of the marks found on the Zaimokuza specimens (Behrensmeyer et al., 1986; Olsen and Shipman, 1988; Fiorillo, 1989; Gaudzinski-Windheuser et al., 2010; Reynard, 2014; Pineda et al., 2014; Rozada et al., 2018).
Previous researchers have put forward the idea that taphonomic processes led to overestimation of the percentage of human-induced traumas on the Zaimokuza remains (Nagaoka et al., 2009; Sakaue, 2016), although they have not directly examined the materials. The original report by Suzuki et al. (1956) only documented the total numbers of individuals and described details of several select specimens with ‘scratch’ marks, so this has neither been confirmed nor denied. Therefore, we aimed to reanalyse the skeletal remains documented in the original report, and investigate the ‘scratch’ marks, considering the implications for the site as a whole.
In the period shortly after excavation, Suzuki reportedly restored 283 crania from the first and second excavations of the Zaimokuza site to use as data for his study. The remaining materials were separated into large wooden boxes and stored in the basement of the Department of Anthropology, The University of Tokyo. Unfortunately, a full list of the specific specimen numbers utilized was not provided in the original report. At present, there are 293 individuals held in skull cases in The University Museum, The University of Tokyo, separated from the larger boxes. The majority are relatively complete skulls, although a small number are only represented by single cranial bones. For 240 of these specimens, original data protocol sheets are present in the museum archive, but these only contain metrics and age and sex determinations, without information regarding analyses of weapon-related trauma. We assumed that 283 of the 293 individuals were the materials that Suzuki utilized in his study and so included all in our present observations. A full list of individuals utilized in this study can be found in the supplementary materials.
In the original report on the Zaimokuza site, Suzuki described individuals using the categories ‘subadults,’ ‘young adults,’ ‘mature adults,’ and ‘old,’ but the exact ages that these categories represent, and the methods of age determination, were not reported. Recent studies on weapon trauma on skeletons from medieval Kamakura mostly differentiate only between subadult and adult individuals (Hirata et al., 2004; Nagaoka et al., 2009, 2010, 2013), so this broad categorization of adult/subadult was comparable and selected in this study. As most of the individuals were represented by crania alone, determination of age was limited to cranial features, and so results should be seen as estimations. The completeness of dental crowns and roots (Ubelaker, 1987), the degree of ectocranial suture closure according to Meindl and Lovejoy (1985), and the development and closure of the occipital synchondrosis (Wakebe, 1990) were utilized, although the latter was not preserved in much of the sample. Individuals displaying full eruption and root formation of the third molars and/or a moderate degree of ectocranial suture closure were classified as adults (approximately 20+ years old), and subadults (approximately 0–19 years old) if not. In cases where permanent dentition was fully erupted but the third molars were not present and little to no ectocranial suture closure was visible, individuals were classified as ‘age unknown.’ All examinations were carried out by the first author.
Sex determination was carried out by the first author on adult individuals using widely recognized cranial features, such as the nuchal crest, mastoid process, supraorbital margin, glabella, and, where present, the mental eminence (Buikstra and Ubelaker, 1994). Sexing of Asian populations using these standards alone is not always accurate, but previous studies have shown that in Japanese populations, sexing using the supraorbital margin had the best accuracy rate (80.5%) followed by the mastoid process (78.6%) (Sakaue and Adachi, 2009). Therefore, in this study if sexual dimorphism was not clearly determinable in all features, but was more pronounced in these two features, individuals were classified accordingly. Indeterminable cases were marked as ‘sex unknown.’ Sex determination was not carried out on individuals whose age was unknown and for subadults as it has been previously described as ‘tentative at best’ (Scheuer and Black, 2004), whilst others have argued that there are no reliable methods for sex determination in subadult individuals (Komar and Buikstra, 2008).
Identifying ‘Scratches’Chapter 7 of the original site report is dedicated to describing the traumas found on the Zaimokuza skeletal remains (Suzuki et al., 1956). According to Suzuki, ‘scratches’ are the shortest, shallowest marks found on the assemblage, measuring less than 10 mm and appearing to the naked eye as a simple line, but when enlarged, appearing more like a very light cut. In contrast to the other two types of sword wounds (‘gashes’ and ‘incisions’) that generally appear independently, ‘scratches’ are mostly clustered together. The direction of each cut is often parallel to each other within the group but in some cases, they cross each other and form facets. They are abundantly found on all bones of the cranial vault, especially on the frontal and parietal bones. Suzuki postulated that they were inflicted as a result of a gradual scraping of a weapon like a sharp Japanese sword at an acute angle to the wall of the cranium (Suzuki et al., 1956). Eight specimens (Nos. 95, 99, 135, 631, 643, 654, 707, and 710) are explicitly mentioned in the text, and the marks they have are described in detail in the original site report with pictures, and these were used as references. In addition, during course of the investigation, a number of the specimens were found with marks circled in pencil (Figure 2). These marks did not display common macroscopic diagnostic features for human-induced cut marks commonly used in recent studies such as linearity, a V-shaped profile, a clean edge, and a flat, smooth polished surface (Boylston, 2000), but were assumed to be marks that Suzuki included as ‘scratches’ and so were counted where present.
Examples of ‘scratches’ circled in pencil directly on specimens: (A) No. 82; (B) No. 454; (C) No. 454; (D) No. 111.
All 293 crania were first examined by the first author macroscopically under incandescent light for any lesions to the bone surface matching Suzuki’s descriptions, and further using head-mounted magnifying glasses at 2.5× magnification. In the present study, ‘one scratch’ denotes the presence of a single mark or a cluster of marks with similar morphologies and directions in close proximity to one another. The location was recorded along with other diagnostic features based on descriptions of ‘scratch’ marks found at the Gokurakuji site in Kamakura (Sakaue, 2016), which was also excavated in the 1959 by Suzuki with identical burial conditions to the Zaimokuza site. Sakaue (2016) identified ‘scratch’ marks on skulls excavated from the Gokurakuji site and proposed that although the marks classified as ‘scratches’ are all similar in that they do not penetrate the external laminae of the skull, they appear to have different morphologies (‘short and sharp,’ ‘weak and slight,’ ‘flat surface with striations,’ and ‘dull and clear’). Therefore, ‘scratches’ found in this study were placed in one of these categories. A full list of the specimens and individual ‘scratches’ identified was compiled along with photographs of each mark.
Modern protocols for the examination of bone surface modifications advise against identification using macroscopic examination alone, and several microscopic features, including the trajectory of the groove and the shape of the apex, are used to distinguish cut marks caused by either stone or metal tools from those caused by natural processes (Andrews and Cook, 1985; Behrensmeyer et al., 1986; Olsen and Shipman, 1988; Domínguez-Rodrigo et al., 2009). Cut marks produced by stone or steel tools in a slicing motion tend to be V-shaped in cross section, with a regular outline that differs from other damage produced either taphonomically or from post-mortem cleaning, sanding, and other modifications, which tends to exhibit U-shaped apexes (Shipman and Rose, 1984; Andrews and Cook, 1985; Behrensmeyer et al., 1986; Olsen and Shipman, 1988; Noe-Nygaard, 1989; Domínguez-Rodrigo et al., 2009). In particular, marks produced by metal weapons have steep and smooth V-shaped profiles (Greenfield, 1999). ‘Scrape’ marks caused by weapons may lack a defined sharp apex but exhibit flat surfaces often with multiple striations caused by irregularities on the sharp edge of the artefact (Binford, 1981; Shipman, 1981). Although flat surfaces and even striations may not be present in scrape marks produced by stone tools, particularly retouched flakes, the human remains from the Zaimokuza site date to the medieval period, and thus it is highly unlikely that stone tools were used. As a large number of marks was found, select marks with different subcategories (though all meeting Suzuki’s criteria for ‘scratches’) were utilized for closer inspection of such microscopic features by scanning electron microscopy (Hitachi Miniscope TM3030, Japan). For each mark, negative impressions were made from silicone-based impression material (Coltene, Affinis Precious Light Body) and examined at between 30× and 100× magnification.
Examination of the crania found that 38.2% of individuals (n = 112/293) displayed marks matching Suzuki’s descriptions of ‘scratches.’ Table 2 shows age and sex determination of the total number of individuals, and those with ‘scratches.’ There is no correlation between the presence of ‘scratches’ and either age or sex, a result that is the same as the original study (Suzuki et al., 1956). The total number of ‘scratches’ identified on specimens numbered 235, with many individuals having more than one group. A list of all the ‘scratches’ found on the individuals with additional information is shown in the supplementary materials. The marks appear most abundantly on the frontal (70/235, 29.8%), and both right (74/235, 31.5%) and left (63/235, 26.8%) parietal bones, as also previously indicated by Suzuki, although they are also found on the occipital (20/235, 8.5%) and on both right (6/235, 2.6%) and left (2/235, 0.9%) temporal bones. The ‘scratches’ themselves vary greatly in their morphologies, with some occurring as single lines, which may be straight or curved, and have a U- or V-shaped cross section. The majority are groups with striations in close proximity to one another, and sometimes parallel to one another. In all except one case the striations have different start and end points and lengths. Many groups have marks that cross over and form facets, often with differing lengths and depths within the same group. Around half of the individuals identified have more than one group of ‘scratches,’ with several individuals displaying up to seven different groups. It should be noted that ‘scratches’ were observed on a lower number of individuals than the original study (38.2% as opposed to 54.1%). However, as described in the Methods section, complete replication of Suzuki’s study was not possible for reasons including the lack of information about the actual specimens utilized and the degree to which his diagnostic standards were met.
| All individuals | % | No. of individuals with ‘scratches’ | % | % of total no. of individuals examined with ‘scratches’ | |
|---|---|---|---|---|---|
| All Adults | 201 | 68.6% | 65 | 58.0% | 22.2% |
| Male | 143 | 71.1% | 42 | 64.6% | 14.3% |
| Female | 45 | 22.4% | 17 | 26.2% | 5.8% |
| Unknown Sex | 13 | 6.5% | 6 | 9.2% | 2.0% |
| Subadult | 55 | 18.8% | 32 | 28.6% | 10.9% |
| Unknown Age | 37 | 12.6% | 15 | 13.4% | 5.1% |
| Total | 293 | — | 112 | — | 38.2% |
When comparing and categorizing the marks observed in this examination according to those found on the Gokurakuji specimens (Sakaue, 2016), 82.3% of the ‘scratches’ identified (194/235) matched the description ‘weak and slight’ and were found on 91.1% of individuals identified in total (102/112). They displayed macroscopically U-shaped cross sections, with the outer edge generally rounded, and included both single marks and groups. Some occurred in patches of macroscopically parallel lines of differing lengths and depths. Others occurred more randomly. Examples include No. 76 (Figure 3), No. 29 (Figure 4), No. 90 (Figure 5), No. 82 (Figure 6), and No. 619 (Figure 7). Specimens shown as examples in the original report including No. 95 (Suzuki et al., 1956: Plate 9a, b), Nos. 99 and 135 (Suzuki et al., 1956: Plate 9c, d), No. 707 (Suzuki et al., 1956: Plate 7a–d), and No. 710 (Suzuki et al., 1956: Plate 8a–c) all have marks matching this description but differ slightly from others as they appear over a larger surface area. However, the individual striations that make them up tend to be shallow, with different start and end points, and have smaller striations running perpendicularly across them as can be seen in No. 710 and No. 95 (Figure 8). Some of the marks in this category display a polished effect but with an uneven surface due to the differing depths of the individual striations. Those categorized as ‘short and sharp’ made up 9.4% of the total amount of marks (22/235) and were found on 14.3% of individuals with ‘scratches’ (16/112). These types of marks were sometimes found near to the deeper ‘incisions.’ A number of these marks more closely resemble shallow ‘incisions’ and some show features such as a albeit small, polished cut surface and flaking on one side, as seen in No. 643 (Suzuki et al., 1956: Plate 11d, e) and No. 654 (Suzuki et al., 1956: Plate 10a–d). A further 6.8% (16/235) matched the description of ‘dull and clear’ and were present on 11.6% of individuals with ‘scratches’ (13/112). Most appeared as single marks, with a similar shape and size to ‘incisions’ but lacking a clear polished cut surface, such as No. 462 (Figure 9). Finally, 1.3% (3/235) matched the description of ‘flat surface with striations’ and were found on 2.7% of individuals identified as having ‘scratches’ (3/112) and included No. 631 previously reported by Suzuki (Suzuki et al., 1956: Plate 6b).
No. 76–single ‘weak and slight scratch.’ The circle indicates the ‘scratch’ mark. The square indicates the area from which the scanning electron microscopy image was taken. Arrows indicate the uneven U-shaped apex of the mark.
No. 29–multiple ‘weak and slight scratches’ around the same vicinity, running in different directions and occasionally crossing over one another. The circle indicates the ‘scratches.’ The square indicates the area from which the scanning electron microscopy image was taken. Arrows indicate striations of varying depths and lengths with U-shaped apexes.
No. 90–multiple ‘weak and slight scratches’ in the same vicinity, but with different depths and lengths. The circle indicates the ‘scratches.’ The square indicates the area from which the scanning electron microscopy image was taken. Arrows indicate striations of varying lengths and depths with uneven U-shaped apexes and weaker striations running perpendicular to the main set of marks.
No. 82–multiple ‘weak and slight scratches’ mostly parallel, but with different depths and lengths. The circle indicates the ‘scratches.’ The square indicates the area from which the scanning electron microscopy image was taken. Arrows indicate striations of varying lengths and depths with uneven U-shaped apexes.
No. 619–multiple ‘weak and slight scratches’ that macroscopically resemble ‘scrape’ marks. The circle indicates the ‘scratches.’ The square indicates the area from which the scanning electron microscopy image was taken. Arrows indicate striations of varying lengths and depths with uneven U-shaped apexes.
Larger patches of ‘weak and slight scratches’ found on No. 710 (left) and No. 95 (right). The individual striations have different lengths and depths and have smaller striations running perpendicularly across them.
No. 462–two ‘dull and clear scratches’ that resemble ‘incisions.’ The circles indicate the ‘scratches.’ The square indicates the area from which the scanning electron microscopy image was taken. Arrow indicates the V-shaped apex.
Various causes of the ‘scratches’ found on medieval Kamakura skeletal remains have been considered previously. According to Suzuki, the marks were assumed to be the result of the scraping of soft tissue from the skull, and in the original Zaimokuza report, they were noted as being found on all individuals who had ‘incisions.’ Suzuki hypothesized that they were caused when Buddhist monks collected the heads and removed the flesh in preparation for funerary rituals because of additional skeletal remains exhumed at the nearby Kuhonji temple by Suzuki and his team just prior to the Zaimokuza excavations. These remains were originally discovered at a site adjacent to the Zaimokuza site in the 1930s and had Buddhist scriptures written on them in black Indian ink reportedly dating to the late Kamakura period, with some individuals displaying ‘scratches’ underneath the black ink (Suzuki et al., 1956). As for those found on individuals from surrounding sites, Nagaoka thought that the ‘scratches’ found on Yuigahama Chūsei Shūdan Bochi No. 372 skeletal remains were likely the result of trophy-taking in battle, judging from the burial conditions of the mass cranial grave and the presence of other weapon-related traumas (Nagaoka et al., 2010). References to the removal of body parts, particularly heads, as trophies of war for reward in medieval Japan are not uncommon (Conlan, 2003; Tsang, 2007). However, Nagaoka also noted that there is no record of scalping in medieval Japan like that found in other parts of the world, and he made some alternative explanations including “primitive types of autopsy, therapy for toothache, or the removal of a scalp before burial as a keepsake” (Nagaoka et al., 2009). Suzuki himself also noted historical references to the usage of human body parts for medicine in the medieval period which may provide a reason for the removal of skin from the scalp unrelated to battle in a later publication (Suzuki, 1989). Finally, Sakaue (2016) deemed that the ‘scratches’ found on the Gokurakuji remains were not caused by the same instrument and reason, depending on their categorization, and considered previous hypotheses as well as damage incurred during the curation process after the excavation.
In the current study, most of the ‘scratch’ marks fell into the category of ‘weak and slight’ and these ‘scratches’ displayed a range of varying morphologies. Clusters that resembled patches such as seen in No. 82 (Figure 6) and No. 619 (Figure 7) were of particular interest as they macroscopically appear like ‘scrape’ marks described in other literature (Binford, 1981; Shipman, 1981). However, if these marks resulted from being scraped with a sharp metal weapon held at an acute angle to the bone, one would expect the scrape to have a flat surface with some striations visible due to damage or serration on the blade itself, with the striations to start and end in the same positions, showing linearity and being parallel throughout. When viewed microscopically, patches such as those found on No. 82 (Figure 6) and No. 619 (Figure 7) have individual striations that differ in length and thickness and begin and end at different points, and so are more like marks caused by sedimentary abrasion found in other literature (Behrensmeyer et al., 1986; Olsen and Shipman, 1988; Fiorillo, 1989; Reynard, 2014). Oliver (1986) noted that marks mimicking scraping were common in trampling, and in particular experiments such as those undertaken by Reynard (2014) and Rozada et al. (2018) have shown that such marks can be produced when bones are trampled in varying types of coastal sediments For other ‘weak and slight scratches’ including those found on No. 76 (Figure 3), No. 29 (Figure 4), and No. 90 (Figure 5), scanning electron microscopy observation showed these marks are mostly flat in cross-section and very shallow, often accompanied by other sets of parallel shallow marks, making them more consistent with marks from sedimentary abrasion produced during trampling too. Studies have shown that repeated or prolonged episodes of trampling can cause initially V-shaped profiles to become more rounded and U-shaped, resulting in varying morphologies of marks in the same vicinity, and on the same individual (Behrensmeyer et al., 1986; Olsen and Shipman, 1988; Domínguez-Rodrigo et al., 2009; Reynard, 2014; Rozada et al., 2018).
Alternatively, Sakaue supposed that some of the ‘weak and slight scratches’ observed on remains from the Zaimokuza and Gokurakuji sites might have been caused by preparation damages during cleaning, restoration, and curation due to them possibly being handled in the same laboratory. However, for the Gokurakuji remains he stated that this was not plausible because the ‘scratches’ did not show fresh breaks even after being rinsed with acetone, implying that damage had occurred before burial. For the Zaimokuza remains, it was not possible to confirm whether all of the marks were peri-mortem as the restored specimens have been covered by a thin coat of preservative, but interestingly one individual (No. 601) displayed marks on the clay used originally to restore the specimen that have a similar morphology to ‘weak and slight scratches.’ The marks do not extend onto the surrounding bones, and so it is plausible to postulate they were made at the time of restoration, but other specimens with similar reconstructions using clay and other materials do not have such marks. It is therefore possible that some of the ‘scratches’ on the Zaimokuza specimens may have been inflicted post-excavation.
Although the human remains from Zaimokuza included 10 individually buried whole skeletons, the majority of the skeletal material from the first and second excavations was grouped into 24 burial pits of varying sizes and depths, some of which consisted almost entirely of human crania not associated with mandibles or postcranial bones. They are thought to have been buried after substantial skeletonization, due to their disarticulation and the presence of carnivore bite marks on a number of the remains (Suzuki et al., 1956). According to Suzuki`s theory, this was because in the aftermath of battle corpses could not be dealt with immediately. However, historical studies have uncovered references to the ill and infirm being abandoned into rivers or on roadways and the corpses being simply thrown away in fields and dry riverbeds (Katsuda, 2003; Suitō, 2009). Although this practice, known as ‘corpse abandonment,’ was traditionally thought as an ancient burial practice, it has been considered to continue to the medieval period (Katsuda, 2003). Suzuki also referred to the early medieval scroll known as Gaki Zoshi which shows corpses in various stages of decomposition, portraying whole skeletons, and disarticulated crania and long bones deposited on the ground with dogs and birds feeding on them (Suzuki et al., 1956). This scroll has been also cited recently as a depiction of medieval ‘corpse abandonment’ (Katsuda, 2003; Suitō, 2009). Taking the Gaki Zoshi scroll into consideration, Uzawa examined taphonomic signatures of carnivore scavenging and animal trampling on mixed human and animal skeletons from the Yuigahama Minami site to trace indications that the corpses were left unburied (Uzawa, 2007). According to him, the bones from mixed human and animal burials from the Yuigahama Minami site are representative of this practice. Archaeological evidence for such a practice in other regions of Japan is a little scarce but in sites such the Dazaifu Joubou in Fukuoka (Yanagi and Yamamura, 2007) and Uramawari in Niigata (Mizusawa, 2007), disarticulated human remains were found discarded in ditches with the remains of horses and other animals. Such practises would have provided ample time for the secondarily buried remains from the Zaimokuza site (which make up the majority) to have been subjected to processes such as trampling before their eventual collection and burial. This is thus a plausible reason for the high frequencies of “‘weak and slight”’ marks found in this study. When considering the hypothesis in relation to the Kuhonji remains with the Buddhist scriptures written over some of the marks (Sakaue, 2016: Figure 40), the black ink may have been applied after such processes took place before their eventual burial. No such scriptures are present on the skeletal remains from the Zaimokuza site examined in this study.
As for the ‘scratches’ in the other three categories, firstly it is plausible that the marks that show a ‘flat surface with striations’ may be the result of a bladed weapon being drawn at an acute angle across the surface of the skull, as indicated by Sakaue (2016), as they show a polished cut surface with linear striations starting and ending at the same points. In addition, ‘short and sharp scratches,’ especially when falling near or by ‘incisions,’ are likely to have been caused by the same weapon as the ‘incisions’ but with less penetrating force as they show linearity and V-shaped profiles, which support Suzuki’s theory that ‘scratches’ and ‘incisions’ are related to each other. However, other ‘scratches’ in this category unaccompanied by ‘incisions’ were identified in this study and reported on crania from Gokurakuji and the Yuigahama Chūsei Shūdan Bochi No. 372 sites. Examples such as those seen on No. 654 and No. 643 (Suzuki et al., 1956: Plates 10 and 11) resemble marks from a sharp instrument being repeatedly scored across the cranium. The reason for this is unknown, but as stated by Sakaue (2016), these types of cut marks cannot be regarded as the result of scalping, but rather some kind of post-mortem mutilation. Finally, the ‘dull and clear’ ‘scratches,’ although lacking a smooth cut surface macroscopically, displayed a linear, V-shaped profile typical of sharp-force weapon trauma when viewed under the scanning electron microscope, such as seen on No. 462 (Figure 9). Taphonomic effects such as trampling can alter features of pre-existing cut marks, eliminating the sharpness and internal striae (Behrensmeyer et al., 1986). The ‘scratches’ in these three categories may therefore be attributed to peri-mortem human activity.
The results of this study indicate that marks displaying different morphologies under the term ‘scratches’ may be attributed to different causes and for different reasons at specific sites as indicated by Sakaue (2016), with some being peri-mortem human-induced marks and others from taphonomic and other post-excavation practises. The revised number for probable peri-mortem human-induced ‘scratch’ marks includes all those with marks matching the descriptions ‘short and sharp,’ ‘flat surface with striations,’ and ‘dull and clear,’ on the Zaimokuza specimens is 9.6% (28/293), which is still the highest compared to surrounding sites. It is likely that the grave pits at medieval Kamakura sites differ in the situations surrounding their deaths, but no patterns between age and sex ratios, frequencies of trauma, and types of burial could be seen in the current study, due to the lack of information on grave numbers for the specimens utilized. Overall, 61 out of 293 individuals (20.8%) displayed sharp-force trauma including ‘gashes,’ ‘incisions,’ and likely human-induced ‘scratches.’ It cannot be ruled out that some of the individuals buried in some of the grave pits at the Zaimokuza site were victims of the siege of Kamakura as previously postulated by Suzuki et al. (1956), but similar traumas found across surrounding sites suggest that incidences of violence were not unusual in the medieval period in Japan. Some of the interments at Zaimokuza may be the result of common burial practices of the period including the practise of ‘corpse abandonment,’ particularly in times of hardship such as during famines and natural disasters where large numbers of people perished in a short interval, and the bodies could not be dealt with swiftly.
The results of this study suggest that the majority of previously reported ‘scratches’ found on the Zaimokuza remains did not result from peri-mortem human-induced violence. Utilizing both macroscopic and microscopic observations, this study found that the majority of marks matching Suzuki’s description of ‘scratches’ are ‘weak and slight,’ and similar in morphology to marks caused as a result of sedimentary abrasion (Behrensmeyer et al., 1986; Olsen and Shipman, 1988; Gaudzinski-Windheuser et al., 2010; Pineda et al., 2014; Reynard, 2014; Rozada et al., 2018). They occur randomly on all elements of the skull, sometimes in patches of macroscopically parallel lines (resembling scraping) but with different lengths and depths, and are often intersected by abundant superficial marks. However, even if the majority of the finer ‘scratches’ are considered as not being human-induced, a portion of the Zaimokuza specimens do display signs of traumatic injury incurred around the time of death. The variability of marks seen in previous studies and observed in this study highlights the difficulty and problems of analysis of bone surface modifications on archaeological specimens, particularly those from large assemblages, and there is currently no completely objective method to accurately classify such marks. However, recent research in the field has introduced methods such as 3D digital microscopy, geometric morphometrics, and deep learning in an attempt to negate the subjective nature of such studies (Courtenay et al., 2020a, 2020b; Domínguez-Rodrigo and Baquedano, 2018; Domínguez-Rodrigo et al., 2020). In the future, these methods should be applied to the marks on the Zaimokuza remains and those from surrounding sites.
We would like to thank Professor Gen Suwa and Professor Yosuke Kaifu, The University Museum, The University of Tokyo, for allowing us to use valuable samples and for their helpful suggestions throughout the course of this research. This research was supported in part by JSPS KAKENHI, numbers 21H04983 and 21K00959.
