Morphological characteristics of buried dog remains excavated from the Kamikuroiwa Rock Shelter site, Ehime Prefecture, Japan

Abstract

In 1962, two sets of dog remains were excavated at the Kamikuroiwa Rock Shelter site in Ehime Prefecture. ¹⁴C age dates of the dog remains correspond to a time period from the end of the Initial Jomon period to the beginning of the Early Jomon period: this indicates that they are the oldest buried dog remains discovered to date in the Japanese archipelago. Both sets of remains represented adult dogs and showed complete permanent dentition. The interred bodies were small, including the bones of their extremities, but they still resembled Jomon dogs of later periods. The muscles had developed, especially those required for mastication, and for the bending and stretching of the extremities. The teeth showed damage due to attrition that most likely occurred prior to death. It has been surmised that the dogs were subjected to extreme levels of stress on their teeth and were buried after tooth loss. Such damage may be related to hunting for large game mammals such as wild boar, similar to Jomon dogs of later periods. Therefore, it is highly possible that these two individuals shared similar characteristics as hunting dogs with Middle, Late, and Final Jomon dogs.

Introduction

At the excavation survey of the Kamikuroiwa Rock Shelter site in Ehime Prefecture in 1962, two sets of buried dog remains were excavated (Esaka et al., 1967). After the excavation, these dog remains went missing without being described in detail. Recently we rediscovered them in a laboratory repository at Keio University, approximately 50 years after their discovery at the Kamikuroiwa Rock Shelter site (Sato et al., 2015). Direct ¹⁴C dating of the dog remains gave c. 7400–7200 calBP (Gakuhari et al., 2015), indicating that these are the oldest buried dog remains discovered to date in the Japanese archipelago.

A great deal of information on the lives of the dogs before their deaths can be ascertained from the buried remains. In this study, we describe their morphological characteristics, and then consider possible uses of dogs in the Initial and the Early Jomon periods, a topic that has been little discussed until now.

Materials and Methods

The dog remains reported in this study were the crania and postcrania of two different sized adult individuals (Table 1, Figure 1, Figure 2, Figure 3, Figure 4). They were rediscovered in a mixed state from a single box, and had been extensively broken into large and small fragments.

Table 1 Inventry of the Kamikuroiwa dog burials

Dog No. 1		Dog No. 2		Dog No. 1/No. 2, indistinctive
KR1001	Cranium	KR2001	Cranium	KR4302	Cervical vertebra
KR1002	Left mandible	KR2002	Left mandible	KR4303–4306	Four thoracic vertebrae
KR1003	Right mandible	KR2003	Right mandible	KR4307–4312	Six lumbar vertebrae
KR1901	Atlas	KR2301	Atlas	KR4313	Sacram
KR1301	Axis	KR2302	Axis	KR4314	Caudal vertebra
KR1302	3rd cervical vertebra	KR2303	3rd cervical vertebra	KR4101	Right ilium
KR1303	5th cervical vertebra	KR2304	4th cervical vertebra	KR4102	Right ischium
KR1304	6th cervical vertebra	KR2305	6th cervical vertebra	KR4103 and 4104	Left iliums (perhaps same individual)
KR1305	7th cervical vertebra	KR2306	7st thoracic vertebra	KR4201–4211	Metacarpals and metatarsals
KR1306	1st thoracic vertebra	KR2108	Right scapula	KR4212	Proximal phalanx
KR1111	Right scapula	KR2109	Left scapula
KR1112	Left xcapula	KR2101a and 2101b	Right humerus
KR1101a and 1101b	Right humerus	KR2102	Right radius
KR1102	Left humerus	KR2103	Left radius
KR1103	Right radius	KR2104	Right ulna
KR1104	Left radius	KR2106	Left ulna
KR1105	Right ulna	KR2105	Right femur
KR1106	Left ulna	KR2107	Left tibia
KR1107	Right femur	KR2201	Right talus
KR1108	Left femur	KR2202	Left talus
KR1109	Right tibia
KR1110	Left tibia
KR1201	Right talus

Figure 1

Superior, lateral, and inferior aspects of crania from Dog No. 1 and Dog No. 2.

Figure 2

Frontal and occlusal aspects of crania from Dog No. 1 and Dog No. 2.

Figure 3

Lateral and occlusal aspects of the mandibles of Dog No. 1 and Dog No. 2.

Figure 4

Axes and limb bones of Dog No. 1 and Dog No. 2.

Water-soluble bonding agents were used to join the broken bones. The relatively small cranium and postcrania were designated as Dog No. 1, and the larger ones as Dog No. 2 (Table 1). However, for several of the remains, such as vertebrae excluding atlas and axis, rib bones, and phalanges, it was difficult to distinguish to which individual they belonged.

Besides our two sets of buried dogs, 24 scattered fragments of dog remains were unearthed from the Kamikuroiwa Rock Shelter (Anezaki et al., 2009). In this study, the relationship between these scattered specimens and the buried dog remains was investigated based mainly on joined fragments.

Measurements were carried out according to the methods of Driesch (1976) and Shigehara (1986). The maxillary and mandibular canine teeth were measured at the tooth cervixes, whereas the incisors were not measured. The measurement values of bones and teeth were rounded to the first decimal place.

Descriptions

Remains of Dog No. 1

Cranium and maxillary teeth

Dog No. 1 was an adult dog with complete permanent dentition. There was considerable damage to the basisphenoid bone periphery and from the basisphenoid bone to the occipital bone. The total length of the cranium (prosthion–akrokranion) was 139 mm, which according to Hasebe (1952) places it in the small class.

The cerebral cranium was low and no pronounced protuberance at the anterior of the frontal bone was observed, compared to modern Japanese dogs. Primitive features remained, such as the side aspect from the snout to the frontal bone, which exhibited a moderate slope, a high zygomatic bone, and also a backward inclination of the line connecting the highest point with the lowest point of the eye socket (Figure 1). In modern small dogs it is normal for the exterior sagittal crest not to have developed. The exterior sagittal angle of Dog No. 1 was practically limited to the interparietal process. However, from the dorsal view the temporal line, which originated from the right and left sides of the interparietal process, did not swell on both sides but instead formed a parabolic line that reached the zygomatic process; this form differed from modern small dogs in Japan. As one part of the temporal line originated from the temporal muscle (m. temporalis) that is involved in closing the mouth (Hermanson and Evans, 1993), it is estimated that the temporal muscle of Dog No. 1, as a small dog, had actually developed considerably. The measurement values for the cranium are shown in Table 2.

Table 2 Cranial measurements (mm)

Driesch No.	Shigehara No.	Measurements	Kamikuroiwa		Early-Final Jomon*
			Dog No. 1	Dog No. 2	n	Mean	SD
1	1	Total length	139	160	25	161	10
2	2	Condylobasal length	139	157	26	151	9
3		Basal length	132	148	1	135
	4	Brain case length	77	81	29	86	7
7		Upper neurocranium length	69	75
8	12	Viscerocranium length	67	82	28	77	7
9		Facial length	76	95
13a	17	Palatal length	71	78	21	78	5
23	10	Greatest mastoid breadth	51	45	34	59	5
25		Greatest breadth of the occipital condyles	28	32
27		Greatest breadth of the foramen magnum	15	18
29	5	Greatest neurocranium breadth	48		34	51	3
30	3	Zygomatic breadth	88		18	87	9
31	8	Least breadth of skull	30	33	33	31	2
32	9	Frontal breadth	38	43	41	41	4
33	11	Least breadth between the orbits	26	30	40	28	3
34	18	Greatest palatal breadth	53	61	28	56	4
	13	Snout length		R = 70, L = 70	29	67	4
36	14	Breadth at the canine alveoli	29		27	32	3
		Nasion–bregma		41
		Inion–basion	39	47

*  Data based on Komiya (1995a, b, 1999, 2002, 2008), Komiya and Tomura (1997), Shigehara (1986), and Shigehara and Onodera (1984).

The non-metric cranial traits (Komiya, 1997) were recorded for 10 items in Dog No. 1 (Table 3). The foramens for the zygomatic nerves opened on both the right and left sides. The rostral alar foramen had not divided into two. There was one foraminulum in the left inner wall of the alar foramen; however, the right inner wall of the alar foramen was lost as a result of damage. The shape of the posterior border of the palatine bone was similar to the b-type of Imaizumi (1980) (Figure 2). The hypoglossal foramens were incomplete on both the left and right sides. The foramen magnum of the occipital bone was circular and of the Onodera et al. (1987)’s A-type.

Table 3 Cranial nonmetric traits

Traits	Dog No. 1		Dog No. 2
	Left	Right	Left	Right
Foramen for zygomatic nerve	+	+	−	−
Intermediate rostral alar foramen	−	−	−	−
Foramen parvum at inner wall of alar foramen	+	/	−	+
Foramen parvum at inner wall of oval foramen	−	−	−	−
Incomplete of hypoglossal canal	+	+	−	−
Dehiscence at ventral condyloid fossa	−	−	−	−
Condyloid canal bridge	−	−	−	−
Accessory condyloid canal	−	−	−	−
Posterior border of palatine: Imaizumi (1980)’s c-type	−		−
Foramen magnum: Onodera et al. (1987)’s A-type	+		/

+, present; −, absent; /, unknown.

Permanent dentition was complete for the maxillary teeth. Some teeth had fallen out on the mesial side from the 2nd premolars (the maxillary 1st, 2nd, 3rd, and 4th premolars are hereafter abbreviated as P¹, P², P³, P⁴), but it is thought that they fell out during the excavation because, excluding the left P², the alveoli were opened. For the left P², where the alveolus was closed, it is thought that the tooth was lost before death or that it was congenitally missing, and there was also degeneration in the corresponding alveolus bone (Table 4). There was wear to the cusps of every type of tooth, from the remaining implanted incisors to the molars, and dentine was also exposed (Figure 2).

Table 4 Maxillary and mandibular dentitions

Dog No. 1

Upper right

Upper left

M2

M1

P4

P3

P2

P1

C

I3

I2

I1

I1

I2

I3

C

P1

P2

P3

P4

M1

M2

A

A

A

A

B

B

A

A

A

A

B

A

B

A

B

C

A

A

A

A

B

A

A

A

A

C

C

A

B

A

/

B

B

B

A

C

C

A

A

A

A

B

M3

M2

M1

P4

P3

P2

P1

C

I3

I2

I1

I1

I2

I3

C

P1

P2

P3

P4

M1

M2

M3

Lower right

Lower left

Dog No. 2

Upper right

Upper left

M2

M1

P4

P3

P2

P1

C

I3

I2

I1

I1

I2

I3

C

P1

P2

P3

P4

M1

M2

/

A

A

A

C

B

A

A

B

B

/

/

A

B

/

A

A

A

A

/

B

A

A

A

A

C

B

/

/

/

/

/

/

/

B

B

C

A

A

A

A

B

M3

M2

M1

P4

P3

P2

P1

C

I3

I2

I1

I1

I2

I3

C

P1

P2

P3

P4

M1

M2

M3

Lower right

Lower left

A, tooth present; B, postmortem tooth loss; C, antemortem tooth loss; /, socket broken.

It is thought that P³ is the first tooth to rotate as the muzzle is shortened (Evans, 1993). For the maxillary tooth row of Dog No. 1, it is thought that there was constriction in the vicinity of P² and that maxillary shortening occurred, therefore it would not be unusual for this sort of misalignment of the teeth to occur. It is possible that rotation of premolars was not seen in this individual because of the antemortem loss of the P² tooth on the left side, meaning that there was room in the premolar row for tooth eruption, and indeed a lingual root was observed in P³ in the right dentition that did not have any missing teeth.

The wear to the maxillary incisors that remained implanted (maxillary 1st, 2nd, and 3rd incisors are hereafter abbreviated as I¹, I², and I³, respectively) was of a medium level, with nodules remaining on the mesial side of both I² and I³. The measurement values of maxillary teeth of Dog No. 1 were within approximately 1 standard deviation (SD) of the mean values for Early–Final Jomon dogs; however, the maxillary tooth row lengths of Dog No. 1 were obviously smaller than the mean values for Early–Final Jomon dogs (Table 5).

Table 5 Measurements of the dentition (mm)

Shigehara No.	Measurements	Kamikuroiwa				Early–Final Jomon*
		Dog No. 1		Dog No. 2
		Left	Right	Left	Right	n	Mean	SD
	Mesiodistal diameter
7	UC	8.0	8.0		9.4	39	8.6	0.8
11	P2			9.1		38	8.4	0.9
13	P3	11.4	11.0		11.4	45	10.1	1.0
15	P4 (maximum)	16.4	16.9	18.1	18.6	62	17.1	1.2
16	P4 (lateral)	16.6	17.0	17.9	17.9	62	16.9	1.1
18	M1	11.5	11.4	12.1	12.1	58	10.9	0.6
21	M2		6.2			52	6	0.6
7	LC	8.8	8.7			47	8.7	0.6
13	P3	9.3	9.8		9.7	61	8.6	0.9
15	P4	10.5	10.7		11.0	77	9.7	0.8
	M1	17.7	18.4	19.4	20.0
17	M1 (maximum)	18.8	18.6	19.7	20.0	89	18.7	1.1
20	M2	8.2	8.3		8.6	71	7.8	0.7
	Buccolingual diameter
8	UC	5.1	5.1		5.7	39	5	0.5
12	P2			3.6		38	3.5	0.3
14	P3	4.5	4.0		4.5	46	4.2	0.4
17	P4	9.1	8.7	9.4	9.2	61	8.7	0.6
19	M1 (maximum)	14.8	14.9	16.5	15.8	38	14.8	0.9
20	M1	13.0	13.0	14.1	13.8	58	13.7	1.4
22	M2		9.4			51	9.2	0.9
8	LC	5.2	5.2			48	5.3	0.4
14	P3	4.1	4.2		4.8	59	4	0.3
16	P4	5.4	5.5		6.0	76	5	0.4
18	M1 (mesial)	7.4	7.3	7.9	7.7	71	7.3	0.4
19	M1 (distal)	6.6	7.1	7.6	7.5	69	7.1	0.4
21	M2	5.4	5.4		5.6	70	6.1	0.5
	Length of the upper tooth row
23	Tooth row L.	71.6	73.3			10	80.3	4.5
25	Premolar row L.	36.2	39.0	46.7	45.9	10	41.9	2.4
	C-P4 row L.	38.0	40.4		47.2	5	45.1	3.3
26	Cheek teeth L.	50.4	50.4			10	53.6	2.2
	C-M3 row L.	50.5	51.0			5	57.3	3.8
27	Molar row L.	17.2	16.5			15	14.7	0.8
30	Dental arch B.	51.7				7	53.4	2.8
	Length of the lower tooth row
25	Premolar row L.			29.1	30.7	8	31.1	2.9
	C-P4 row L.	31.3	31.5	35.8		8	35.1	3.2
26	Cheek teeth L.			60.1	62.3	6	59.1	3.7
	C-M3 row L.	60.4	60.6	66.4		6	65.7	4.8
27	Molar row L.	29.6	30.2	30.6	33.0	8	29.9	2.0

*  Data based on Komiya (1995a, b, 1999, 2002, 2008), Komiya and Tomura (1997), Shigehara (1986), and Shigehara and Onodera (1984).

Mandibles and mandibular teeth

For the left and right mandibles that were compatible with the crania, both sides were damaged in the vicinity of the canine tooth alveolar mesial edge and restoration of the infradentale was difficult. There was damage to the left mandible on the surface of the coronoid process and to the right mandible at the base of the angular process. Both pieces of damage appeared to be new and it was estimated that they occurred post-excavation. The body of the mandible was thick compared with modern Japanese dogs. The masseteric fossa, which is the end part of the masseter muscle, was deep, and the coronoid crest and condyloid crest were clearly visible (Figure 3). The measurement values for the mandible are shown in Table 6.

Table 6 Mandibular measurements (mm)

Driesch No.	Shigehara No.	Measurements	Kamikuroiwa				Early–Final Jomon*
			Dog No. 1		Dog No. 2
			Left	Right	Left	Right	n	Mean	SD
4		Length: the condyle process–aboral border of the canine alveolus	91
5		Length from the indentation between the condyle process and the angular process–aboral border of the canine alveolus	88
6		Length: the angular process–aboral border of the canine alveolus	91
	22	Breadth of the mandibular ramus	27		29	29	5	28	2
18		Height of the vertical ramus	41
19		Height of the mandible behind M1	18	18	22	21
	23	Mandibular height-1	19	19	22	22	72	21	2
	24	Mandibular height-2	18	18	22	22	81	21	2
	25	Mandibular height-3	16	17	20	20	69	19	2
	26	Mandibular thickness	9	9	11	11	81	10	1
	27	Depth of the masseter fossa	8		8	8	36	7	1

*  Data based on Komiya (1995a, b, 1999, 2002, 2008), Komiya and Tomura (1997), Shigehara (1986), and Shigehara and Onodera (1984).

In the left mandible, five teeth, namely the canine tooth and the teeth from the 3rd premolar to the 2nd molar, remained implanted (the 1st, 2nd, 3rd, and 4th mandibular premolars are hereafter abbreviated as P¹, P², P³, and P⁴, and the 1st, 2nd, and 3rd molars are abbreviated as M₁, M₂, and M₃) and were slightly crowded. Six teeth remained implanted in the right mandible: the 1st incisor, the canine tooth, and teeth from P₃ to M₂. The M₃ alveoli were opened on both the left and right sides and thus it is possible that the tooth fell out during the excavation. The two alveoli between P₃ and the canine teeth were in the process of closing on both the left and right sides, but the broken root of a tooth remained within the alveolus on the right distal side. There are a number of possible interpretations for which teeth these two alveoli originated from, but it is most likely that both the left and right alveoli originated from P₂, which fell out before death, and P₁ is considered to be the missing tooth on both the left and right sides (Table 4). Wear could be seen in every type of tooth that remained implanted, with protoconid wear being particularly severe for the left M₁ and paraconid wear being more noticeable than protoconid for the right M₁. The wear level of M₁ corresponded to that of the maxillary P⁴ and the maxillary 1st molar (the maxillary 1st and 2nd molars are abbreviated to M¹ and M², respectively), which might be associated with frequent use of the upper and lower carnassial teeth before death and a lifestyle involving the biting and breaking of hard objects. The measurement values for the mandibular teeth are shown in Table 5.

Long bones of the extremities

The left and right humerus, radius, ulna, femur distal part, and tibia were identified (Figure 4). All of the epiphysis was fused and was compatible with observations of the cranium in the growth stage. Because observations were limited by the range of bones that could be restored, abnormalities such as fractures and diseases were not observed. Compared to modern dogs in Japan, the deltoid tuberosity of the humerus was wide and the tuberous sculpting was strong. The tuberosity for the teres major was well developed and the indentation behind the tricipital line of the humerus neck to the head was deep. As the muscles that act in the flexing and extension of the shoulder joint and elbow joint, such as the m. deltoideus, the m. teres major, the m. brachialis, and the m. triceps brachii, all originate or end in these tuberosities and the humerus proximal-side surface part (Hermanson and Evans, 1993), it was understood that these muscles were considerably developed in this individual compared with modern dogs. Furthermore, the development of the muscle attachments in the long bones in the extremities other than the humerus was excellent and it is estimated that the dog had superior mobility in the bending and stretching of the extremities and in the bending digits.

The measurement values are shown in Table 7. In addition, deviation graphs of well-preserved radii and tibiae are shown in Figure 5 and Figure 6 in order to examine the proportions of the extremities. As the sex of Dog No. 1 from the Kamikuroiwa Rock Shelter is unknown, the horizontal center lines of Figure 5 and Figure 6 are drawn using the mean of male and female Shiba dogs. In both figures, radii and tibiae of Late Jomon buried dogs unearthed from the Tagara shell midden that have large sample numbers are used for comparison (Shigehara and Onodera, 1984).

Table 7 Limb bone measurements

Bone	Driesch Code	Shigehara No.	Measurements	Kamikuroiwa				Early–Final Jomon*
				Dog No. 1		Dog No. 2
				Left	Right	Left	Right	n	Mean	SD
Humerus	GL	6	Humerus length	115				32	124	12
	Dp	7	Proximal maximum breadth	31	31			13	23	4
	SD		Minimum breadth of diaphysis	9			10	2	9
	Bd	10	Maximum breadth in the distal end	25	25			24	26	3
Radius	GL	11	Radius length	106	107		124	29	117	10
	Bp	12	Maximum breadth in the proximal end	14	14			30	15	1
		13	Maximum diameter in the proximal end	9	9		10	11	9	1
		14	Minimum neck breadh	10	10			12	11	1
	SD		Minimum breadth of diaphysis	8	8	9	10	2	9
		15	Breadth in the middle	9	9	11	11	34	10	1
		16	Sagittal diameter in the middle	5	5	6	6	34	7	1
	Bd	17	Maximum breadth in the distal end	18	18	21	20	31	19	2
		18	Maximum sagittal diameter in the distal end	10	10	11	11	11	11	1
Ulna	DPA		Depth of olecranon		18		20
	SDO		Minimum depth of olecranon		16			2	16
	BPC		Breadth across the coronoid process		14
		20	Sagittal diameter in the incisura		10		11	17	11	1
Femur	Bp		Breadth of the proximal end				33	2	28
	DC		Depth of the caput femoris				17	2	14
	SD		Minimum breadth of diaphysis				11	2	9
	Bd	28	Maximum breadth in the distal end	25	25			33	26	2
Tibia	GL	29	Tibia length		119			31	133	14
	Bp	30	Maximum breadth in the proximal end	27	27			29	28	2
		31	Sagittal sagittal diameter in the proximal end	28	28			23	28	5
	SD		Minimum breadth of diaphysis		9	10		2	9
		32	Breadth in the middle	9	9			36	10	1
		33	Sagittal diameter in the middle	9	9			35	10	1
	Bd	34	Maximum breadth in the distal end	17	17	19		32	19	2

*  Data based on Komiya (1995a, b, 1999, 2002, 2008), Komiya and Tomura (1997), Shigehara (1986), and Shigehara and Onodera (1984).

Figure 5

Graphs showing deviation of the limb bone measurements for Dog No. 1 and Dog No. 2 of the Kamikuroiwa Rock Shelter site and male Late Jomon dogs of the Tagara shell midden. Line M: mean of the data for male Shiba dogs. Data of the Tagara Jomon and Shiba dogs were based on Shigehara and Onodera (1984).

Figure 6

Graphs showing deviation of limb bone measurements in the Dog No. 1 and Dog No. 2 of the Kamikuroiwa Rock Shelter site and female Late Jomon dogs of the Tagara shell midden. Line M: mean of the data for female Shiba dogs. Data of the Tagara Jomon and Shiba dogs were based on Shigehara and Onodera (1984).

Although in general the radii and tibiae of Dog No. 1 from the Kamikuroiwa Rock Shelter are smaller than those from the Tagara shell midden, their deviation line patterns are very similar. Proportion is less affected by environment compared to body size, so proportion is considered to be genetically controlled (Eveleth and Tanner, 1990; Holliday, 1997). In this regard, it is noteworthy that limb bone proportions are similar between Dog No. 1 from the Kamikuroiwa Rock Shelter and dogs from the Tagara shell midden, even though they belong to different periods and were unearthed in different locations. This might suggest that the genetic distance between them is not that great.

Remains of Dog No. 2

Cranium and maxillary teeth

Dog No. 2 was also an adult dog with complete permanent dentition. Except for a part of the left maxilla, we were able to connect the cranium almost sequentially. In the side view, the contour line from the frontal bone to the snout was slightly more curved than that of Dog No. 1, but as with Dog No. 1 the zygomatic bone was high and the line joining the highest and lowest part of the eye socket inclined backwards. The cranium total length could be measured and was found to be approximately 160 mm, placing it into the small-and medium-sized class of Hasebe (1952). The sagittal crest was clear. Although the rear part was damaged and missing, the front extended forward to the parietointerparietal suture. Furthermore, on the dorsal view the temporal line reached the zygomatic process while describing a strong parabola (Figure 1).

In the ventral view the palate presented a pear shape that curved in the vicinity of P², and it is thought that shrinkage of the snout was occurring, as was the case with Dog No. 1 (Figure 2). There were missing regions in the areas of the left I¹ and I² periphery, left P¹, left and right M², and M³ periphery, but as tooth eruption occurred in P¹, P², and P³ it was considered that permanent dentition had been completed. As porous alveolus closure markings were observed for the right P², it is thought that this tooth fell out before death (Table 4). The teeth that remained implanted had all suffered attrition, with paracone and metacone attrition being particularly noticeable for P⁴ and M¹ (Figure 2). The shape of the occipital bone foramen magnum was unclear due to damage. The measurement values for the cranium and maxillary teeth are shown in Table 5.

Non-metric cranial traits in the cranium of Dog No. 2 were observed for the nine items shown in Table 3. The posterior border of the palatine bone was partially damaged on the right side, but the restored shape that was reversed on the median line corresponded to the b type described by Imaizumi (1980).

On the basis of comparison of the cranium total lengths and greatest palatal breadths of Dog No. 1 and Dog No. 2 with the measurements of previously reported Jomon adult dog crania from the Early Jomon period onwards and those from the Higashimyo site in Saga prefecture (Initial Jomon period), the straight line in Figure 7 shows the regression line of Jomon dog crania since the Early Jomon period. From this figure, we can see that Dog No. 1 is the smallest in the Jomon dogs group, and also that, although the total length of the cranium of Dog No. 2 was approximately in the middle of the group of Jomon dogs, the greatest palatal breadth was relatively wide. All crania excavated from the Higashimyo site (No. 516, No. 9596, No. 15537; Maruyama et al., 2009) had relatively long total lengths and wide greatest palatal breadths relative to the measurements for other Jomon dogs.

Figure 7

Scatter plots of the total length and the greatest palatal breadth of the cranium in Dog No. 1 and Dog No. 2 of the Kamikuroiwa Rock Shelter site (●), the Higashimyo Initial Jomon dogs (○), and the Early–Final Jomon dogs (×). Data for the Higashimyo dogs were measured by one of authors (H.K.), and data of the Early–Final Jomon dogs were based on Komiya (1995a, b, 1999, 2002, 2008), Komiya and Tomura (1997), Shigehara (1986), and Shigehara and Onodera (1984).

Mandible and mandibular teeth

The left and right mandibles were compatible with the joints and points of engagement with the cranium. All tips of the bodies of the mandibles were damaged in the vicinity of the canine teeth alveoli and restoring the infradentale was difficult. Apart from damage to the upper part of the coronoid process in the left mandible, no major damage to the back part of the mandible was observed (Figure 3). All of the damage was considered to be new, apart from that at the tip of the right mandible.

The left and right P₃, P₄, M₁, and M₂ remained implanted. The M₃ alveolus was open. If we assume that the one alveolus that opened distally among the left and right canine teeth alveoli originated from P₁, it is reasonable to conclude that the traces of porosity between this alveolus and P₃ indicate that P₂ fell out before death. If we accept this to be the case, it can be considered that P₂ of this dog was lost before death or was congenitally missing. P₃ and P₄ on both the left and right were crowded and the teeth that remained implanted were worn. The wear was particularly hard for the M₁ paraconid and protoconid, on the distal teeth cusps that comprise the talonid, and M₂. These wear levels corresponded closely to the maxillary P⁴ and M¹. As with Dog No. 1, it is considered that Dog No. 2 habitually chewed on and broke hard objects during its life.

The left alveolus process between the left M₁ and M₂ was approximately 5.5 mm on the buccal side and had retreated by a depth of approximately 3 mm; the upper part of the M₁ distal root region and the M₂ mesial root were exposed on the buccal side and the tongue side. The possibility that the alveolus process was absorbed as a result of gum disease before death should be considered. The measurement values for the mandibles and the mandibular teeth are shown in Table 5 and Table 6.

Long bones of the extremities

There was considerable damage to the long bones of the extremities of Dog No. 2 and the only bones that could be identified were the right humerus, the left and right radii, ulnae, the right femur, and the left tibia (Figure 4). As with the bones of Dog No. 1, the entire epiphysial line was agglutinated and identified as adult dog bones. No breaks or other abnormalities were observed in the bones that could be recovered. Measurement of total length was only possible in the radii. The total length of the radius of Dog No. 2 was within 1 SD of the mean value for the radii of Early–Final Jomon dogs (Table 7).

Comparison with materials previously reported in 2009

Apart from the buried dog remains analyzed in this research, there are other previously reported dog remains excavated from the Kamikuroiwa Rock Shelter site for which the site name and excavation layer are clearly identified (Anezaki et al., 2009). As described above, the remains of Dog No. 1 and Dog No. 2 were composed of the cranium and bones of the extremities, vertebrae, and rib bones, and other postcranium bones from two adult dogs of different sizes with completed permanent dentition. Vertebrae were confirmed as being one atlas (KR2301) and two axes (KR1301 and KR2302). As it was possible to join the KR2301 atlas vertebra to the occipital condyle of Dog No. 2 without forcing, it was assumed that it belonged to Dog No. 2. As the two axes were of different sizes, it was assumed that one belonged to Dog No. 1 and the other to Dog No. 2, and individual samples of each were collected for stable isotope analysis and ¹⁴C dating.

No atlas vertebrae for Dog No. 1 could be identified from among the excavated dog remains. Also, as the collection of limb bones for Dog No. 2 could not be described as particularly good, it was considered necessary to examine the possibility that some of these excavated dog bones had been mixed up with previously reported materials.

When we examined in detail the duplications and relationships of the sizes and parts of these excavated dog bones with those already reported, we discovered one atlas vertebra (KR1901) that joined with the occipital condyle of Dog No. 1 in addition to radius and femur bones of at least two adult dogs other than Dog No. 1 and Dog No. 2. The epiphyses were fused in both cases. Furthermore, it was confirmed that the materials included one mandible fragment in which the M₁ had not erupted in the body of the mandible and was therefore determined to be from an infant dog, and two separated upper and lower canine teeth that had different levels of attrition than those observed in Dog No. 1 and Dog No. 2.

On the basis of the above findings, it is apparent that the dog remains excavated at the Kamikuroiwa Rock Shelter site were comprised, at the very least, of five individuals, with the bones of two adult dogs and one infant dog in addition to the excavated dog bones from the two individuals reported in this paper. Furthermore, the fact that the atlas vertebra KR1901 joined to the bones of Dog No. 1 supports the notion that some of these excavated dog remains had been mixed in with the already reported materials.

Discussion

In this section we will explain why the excavated dog remains from the Kamikuroiwa Rock Shelter site are thought to be the oldest found in Japan and then consider the archaeological significance of their morphological characteristics.

The oldest buried dog remains in Japan

The ¹⁴C ages of the dog remains were estimated 6442 ± 56 BP [7422–7322 cal BP (68.2%) (1σ) ] for Dog No. 1 and 6359 ± 51 BP [7413–7392 cal BP (8.7%), 7372–7357 cal BP (5.4%) and 7331–7249 cal BP (51.4%) (1σ)] for Dog No. 2 (Gakuhari et al., 2015). These dates correspond to a time period from the end of the Initial Jomon period to the beginning of the Early Jomon period.

Buried dogs of the Initial Jomon period are unknown. Those of the Early Jomon period were discovered at the Hanazumi shell mound in Saitama Prefecture (Saito, 1939), the Nishinoyato shell mound in Kanagawa Prefecture (Sakazume and Takeshita, 1936), the Ugasaki shell mound in Miyagi Prefecture (Shigehara et al., 1980), the Ohguruwa shell mound in Aichi Prefecture (Kaneko, 1983), and the Ushuku Primary School site in Kagoshima Prefecture (Shigehara and Doi, 2003). They belong to the middle or late Early Jomon period (Shigehara et al., 1980; Miyagi Prefecture Board of Education, 1980; Shigehara and Doi, 2003). Therefore, at present, the dog remains from the Kamikuroiwa Rock Shelter site would be the oldest buried dogs with whole bodies in the Japanese archipelago.

Scattered dog remains of the Initial and Early Jomon period

The oldest dog remains found in Japan to date are a fragment of a right mandible excavated at the Natsushima shell mound in Kanagawa Prefecture (Naora, 1973) and a left maxillary canine tooth excavated at the Tochihara cave site in Nagano Prefecture (Miyao et al., 1987). Both were isolated remains and were estimated to be from the middle of the Initial Jomon period.

The late Initial and Early Jomon sites in many parts of the archipelago yielded dog remains, such as at the Tentozan shell mound in Hokkaido (Hatakeyama, 1967), the Akamido shell mound in Aomori Prefecture (Kobayashi, 1989), the Miyanohara site in Kanagawa Prefecture (Kaneko, 1972), the Mibiki site in Ishikawa Prefecture (Shigehara et al, 2004), the Mawaki site in Ishikawa Prefecture (Miyazaki and Hiraguchi, 1986), the Mazukari shell mound in Aichi Prefecture (Watanabe, 1980), the Megumi site in Tottori Prefecture (Inoue, 1986), the Taishakukyo site group in Hiroshima Prefecture (Kawamura, 1992), the Higashimyo site in Saga Prefecture (Maruyama et al., 2009), and the Takashima submarine site in Nagasaki Prefecture (Kimura, 1993). All confirmed findings were of isolated or scattered remains.

As the ¹⁴C ages of these dog remains have not been measured, the above ages may be re-examined. However, with respect to the Higashimyo site, the beginning and the end of the deposition of shells at the site have been calculated as 8000–7850 cal BP and 7800–7700 cal BP, respectively (Nakamura, 2009). Therefore we can certainly estimate the age of the Higashimyo site dog remains; they would be closest in ¹⁴C age to the Kamikuroiwa Rock Shelter site Dog No. 1 and Dog No. 2. Here, it should be noted that there are marked differences in cranium size between the dogs from the Kamikuroiwa Rock Shelter site and the Higashimyo site. The size distribution shown in Figure 7 suggests that different types of dogs existed in the Japanese archipelago and it is very interesting to consider their phylogenetic backgrounds.

Morphological characteristics and role of the dogs buried at the Kamikuroiwa Rock Shelter site

The main morphological characteristics of Jomon period dogs that have been previously identified (Saito, 1936; Naora, 1973; Kaneko, 1978; Nishimoto, 1983; Shigehara and Onodera, 1984; Shigehara, 1988, 1990; Shigehara and Hongo, 2000) are as follows: (1) the body type was a small-sized dog similar to the Shiba Inu class, but the concave shape from the tip of the snout to the head was small; (2) the bones of the extremities were strong; (3) there have been few excavations of individuals with fractured limb bones; (4) the mastication muscles were well developed; (5) there was no misaligned dentition of maxillary teeth, which develop well in modern Japanese dogs; and (6) it is thought that the maxillary and mandibular teeth were used for hunting and antemortem tooth loss is frequently observed. These morphological characteristics were pointed out on the basis of a large number of buried dog individuals of the Middle and Late Jomon periods. Interestingly, both of the two buried dogs from the Kamikuroiwa Rock Shelter site also had many of these features.

The role of dogs within Jomon society, which relied on hunting and gathering for its subsistence, is generally considered to be as hunting dogs. The damage to the teeth, which was assumed to be due to the strong stress placed on the teeth, is frequently observed in Jomon dogs; therefore some have argued that Jomon dogs would have been used for hunting large terrestrial mammals, especially wild boar (Komiya, 1989, 2002; Shigehara et al., 2004). Even for modern-day hunters with guns, hunting wild boar, which is relatively quick and ferocious, tends to be unsuccessful without extremely nimble and suitably trained dogs (Ishikawa, 1971; Ue, 1991; Shirai, 1992).

At the Kamikuroiwa Rock Shelter site, the proportion of deer and wild boar in the vertebrate remains is high (Anezaki, et al., 2009); when considering the morphological characteristics of the remains of Dog No. 1 and Dog No. 2, it seems possible that they were used as hunting dogs and buried in the same cave after death. However, we believe that further comparative studies will be required to address the question of whether the culture of dog breeding at the Kamikuroiwa Rock Shelter site was the same as that at various other sites across Japan from the Middle Jomon period onwards; for example, comparing the hunting periods and the hunting tools used at the Kamikuroiwa Rock Shelter site with those at Middle and the Late Jomon period sites.

Summary

In this research, the authors attempted to extrapolate the uses of Jomon dogs during the Early Jomon period based on the morphological characteristics of two sets of buried dog remains excavated from the Kamikuroiwa Rock Shelter site in Ehime Prefecture, which are the oldest discovered to date in Japan. Both sets of remains were adult dogs with completed permanent dentition, and although their cranium total lengths indicate they were small for Jomon dogs of that period, the size and proportion of their total body, including the bones of the extremities, were consistent with dog bones of the Middle Jomon period onwards. In both sets, the muscles involved in mastication and in bending and stretching the extremities were developed, but there was considerable attrition to their teeth and it was confirmed that this damage occurred before death. Excavated dog remains from various sites from the Jomon period onwards frequently show this type of damage. It is reasonable to conclude that this characteristic was caused by grappling with large game mammals such as wild boar, an activity that placed considerable stress on their teeth. It is considered that these two individuals were hunting dogs, as were Jomon dogs from the Middle Jomon period onwards, and that they were buried in the same cave after death.

Acknowledgments

The authors would like to express their sincere gratitude to Nobuo Shigehara (Nara National Research Institute for Cultural Properties) for his invaluable advice and to Iwao Nishida (Saga City Board of Education), Akira Matsui (Nara National Research Institute for Cultural Properties), Tomoko Anezaki (Gunma Museum of Natural History), and Toyohiro Nishimoto (Date City Institute of Funkawan Culture) for their support and assistance. This research was supported by the following funds: 2012 Keio University Matsunaga Memorial Fund for the Research of Cultural Assets (Title: Research of Buried Dog Remains in the Initial Jomon Period, Research Representative: T. Sato); 2013–2014 JSPS Core to Core Program (Advanced Research Networks) (Title: Advanced Core Research Center for the History of Human Ecology in the North, Research Representative: H. Kato).

References

•  Anezaki  T.,  Yoshinaga  A.,  Sato  T. and  Nishimoto  T. (2009) Vertebrate remains. Bulletin of the National Museum of Japanese History, 154: 325–342 (in Japanese).
•  von den Driesch  A. (1976) A Guide to the Measurement of Animal Bones from Archaeological Sites. Peabody Museum Bulletin 1, Cambridge MA, Harvard University.
•  Esaka  T.,  Okamoto  K. and  Nishida  S. (1967) The Rock Shelter of Kamikuroiwa, Ehime Prefecture. In: Cave Site Research Committee of the Japanese Archaeological Association (ed.), Cave Sites in Japan. Heibonsha, Tokyo, pp. 224–236 (in Japanese).
•  Evans  H.E. (ed.) (1993) Miller’s Anatomy of the Dog, 3rd edn. W.B. Saunders, Philadelphia.
•  Eveleth  P.B. and  Tanner  J.M. (1990) World Wide Variation in Human Growth, 2nd edn. Cambridge University Press, Cambridge.
•  Gakuhari  T.,  Komiya  H.,  Sawada  J.,  Anezaki  T.,  Sato  T.,  Kobayashi  K.,  Ito  S.,  Kyobayashi  K.,  Matsuzaki  H.,  Yoshida  K. and  Yoneda  M. (2015) Radiocarbon dating of one human and two dog burials from the Kamikuroiwa rock shelter site, Ehime Prefecture. Anthropological Science, 123: 87–94.
•  Hasebe  K. (1952) Dog remains. In: Conservation of Cultural Assets Committee (ed.), First Report on the Excavation of the Yoshigo Shell Mound. Conservation of Cultural Assets Committee, Tokyo (in Japanese).
•  Hatakeyama  S. (1967) The early dog remains from Tentozan shell mound, Hokkaido. Hokkaido Archaeological Journal, 3: 21–30 (in Japanese).
•  Hermanson  J.W. and  Evans  H.E. (1993) The muscular system. In  Evans  H.E. (ed.), Miller’s Anatomy of the Dog, 3rd edn. W.B. Saunders, Philadelphia.
•  Holliday  T.W. (1997) Body proportions in Late Pleistocene Europe and modern human origins. Journal of Human Evolution, 32: 423–447.
•  Imaizumi  Y. (1980) Systematic status of the domestic dog in the Family Canidae. Report of the Society for Researches on Native Livestock, 9: 7–52 (in Japanese).
•  Inoue  T. (1986) Faunal remains from the Megumi site. In: Archaeological Excavations Report Associated with the Kamo River Improvement Work. Yonago City Board of Education and River Construction Section of Tottori Prefecture, Yonago, pp. 129–138 (in Japanese).
•  Ishikawa  B. (1971) Actual Hunting Shikoku Dogs. Association for the Preservation of Actual Shikoku Hunting Dog, Hisai (in Japanese).
•  Kaneko  H. (1972) Animal remains from the Miyanohara shell mound. In: Musashino Art University Archaeology Research Group (ed.), Miyanohara Shell Mound. Seiyu-Syobo, Tokyo, pp. 36–45 (in Japanese).
•  Kaneko  H. (1978) Faunal remains excavated at Jomon sites (2). Archaeological Notes, 7: 1–18 (in Japanese).
•  Kaneko  H. (1983) Dog was a good friend of human: dog and human relationship in the Neolithic Jomon period. Anima, 121: 6–11 (in Japanese).
•  Kawamura  Y. (1992) Stratigraphic distribution of mammals in the Taishaku-kyo sites, Hiroshima Prefecture, west Japan. The Quaternary Research, 31: 1–12 (in Japanese with English abstract).
•  Kimura  I. (1993) Faunal remains of the Takashima underwater site, Nagasaki Prefecture. In: Takashima Town Board of Education (ed.), Takashima Underwater Site II.—Emergency Excavations Survey Report in conjunction with the Tokonami Port Repair Works in Takashima Town, Kitamatsura-gun, Nagasaki Prefecture. Takashima Town Board of Education, Nagasaki Prefecture, Takashima, pp. 99–104 (in Japanese).
•  Kobayashi  K. (1989) Faunal remains of the Akamido site. In: Hachinohe City Board of Education (ed.), The Excavation of the Akamido Site. Hachinoe City Board of Education, Hachinoe, pp. 115–122 (in Japanese).
•  Komiya  H. (1989) Dogs of the early Jomon Period in Chiba Prefecture. Bulletin of the Natural History Museum and Institute, Chiba, 1: 1–21 (in Japanese).
•  Komiya  H. (1995a) Re-examination of a Jomon dog from the Sanganji shell midden, Fukushima Prefecture. Bulletin of the Natural History Museum and Institute, Chiba, 4: 27–43 (in Japanese).
•  Komiya  H. (1995b) Buried dog remains from the Kurigasawa site in Matsudo City. Bulletin of the Matsudo Museum, 2: 3–17 (in Japanese).
•  Komiya  H. (1997) Nonmetric cranial variants of modern dogs and the origins of domesticated dogs in Japan. Anthropological Science, 105: 69.
•  Komiya  H. (1999) Jomon dog remains from the Ipponmatsu site in Kamagaya City. Research of the Kamagaya City History, 12: 1–16 (in Japanese).
•  Komiya  H. (2002) The archaeological meaning of the buried dog excavated from the bottom of the flask shaped pit at the Futatsumori shell mound Aomori Prefecture. Bulletin of the Natural History Museum and Institute, Chiba, 7: 37–58 (in Japanese).
•  Komiya  H. (2008) Buried dogs and infant wild boar excavated from the No. 378 pit dwelling. In: Chiba Educational Promotion Foundation Center (ed.), Ariyoshiminami shell mound. Chiba Educational Promotion Foundation Center and Urban Renaissance Agency Chiba Region Branch, Yotsukaido, pp. 387–391 (in Japanese).
•  Komiya  H. and  Tomura  M. (1997) Jomon dog remains from the Sakai site, Chiba Prefecture. Bulletin of the Natural History Museum and Institute, Chiba, 5: 1–17 (in Japanese).
•  Maruyama  M.,  Nagai  R. and  Matsui  A. (2009) Faunal remains. In:  Nishida  I. (ed.), Higashimyo Sites, Vol. 2. The Saga City Board of Education, Saga, pp. 146–178 (in Japanese).
• Miyagi Prefecture Board of Education (ed.) (1980) Miyagi Prefecture Cultural Assets Survey Report 67: Kongoji Shell Mound, Ugasaki Shell Mound, Ugasaki No. 1 Tumulus. Miyagi Prefecture Board of Education, Sendai (in Japanese).
•  Miyao  T.,  Nishizawa  T. and  Hanamura  H. (1987) Mammalian remains of the initial Jomon period at the Rockshelter site of Tochihara, Nagano Prefecture, Japan: 6th report on dog and medium- and small-sized carnivores. Nagano Prefecture Archaeological Journal, 53: 24–38 (in Japanese).
•  Miyazaki  N. and  Hiraguchi  T. (1986) Faunal remains. In:  Yamada  Y. (ed.), The Mawaki Site. The Noto Town Board of Education, Noto, pp. 346–382 (in Japanese).
•  Nakamura  T. (2009) Dating and analysis of ages of deposits. In:  Nishida  I. (ed.), Higashimyo Sites, Vol. 2. The Saga City Board of Education, Saga, pp. 9–15 (in Japanese).
•  Naora  N. (1973) Remains of domestic animals excavated from ancient sites. Benefit Society of the Japan Racing Association, Tokyo (in Japanese).
•  Nishimoto  T. (1983) Dogs. In:  Kato  S.,  Kobayashi  T. and  Fujimoto  T. (eds.), Research on Jomon Culture 2: Subsistence. Yuzankaku, Tokyo, pp. 161–170 (in Japanese).
•  Onodera  S.,  Shigehara  N. and  Eto  M. (1987) Discriminant analysis of the sexual differences of the skeletons in Shiba Inu (Canis familiaris). Journal of the Japanese Association of Anatomists, 62: 19–32 (in Japanese).
•  Saito  H. (1936) Ancestors of Japanese dogs. Nihon-ken, 5: 1–7 (in Japanese).
•  Saito  H. (1939) A report on the domestic dog remains of the Neolithic Japan in the Oyama Institute of Prehistory, and a classification of domestic dog remains of the prehistoric Japan. Journal of Prehistory, 12: 151–263 (in Japanese).
•  Sakazume  N. and  Takeshita  J. (1936) Buried dog skeleton of the Nishinoyato shell mound, Aza-Yamada, Nakagawa Village, Tsuzuki-gun, Kanagawa Prefecture. Journal of Prehistory, 8: 77–93 (in Japanese).
•  Sato  T.,  Hashimoto  M.,  Abe  Y. and  Ando  H. (2015) Rediscovery of the oldest dog burial remains in Japan. Anthropological Science, 123: 99–105.
•  Shigehara  N. (1986) Catalogue of the ancient and recent canid skeletons collected by Dr. Kotondo Hasebe and preserved in the University Museum, the University of Tokyo. The University Museum, the University of Tokyo, Material Report No. 13, Tokyo (in Japanese).
•  Shigehara  N. (1988) Dogs have been man’s friend since ancient times. Anima, 189: 98–101 (in Japanese).
•  Shigehara  N. (1990) Shape changes by era in the Japanese dog. Bulletin of the National Museum of Japanese History, 29: 89–108 (in Japanese with English abstract).
•  Shigehara  N. and  Doi  N. (2003) Domestic Jomon dog excavated from Ushuku Primary School, Kasari-town, Kagoshima Prefecture. Amami Archaeology, 5: 53–60 (in Japanese).
•  Shigehara  N. and  Hongo  H. (2000) Ancient remains of Jomon dogs from Neolithic sites in Japan. In:  Crockford  S.J. (ed.), Dogs Through Time: An Archaeology Perspective. BAR International Series 889, Archaeopress Publishers of British Archaeological Reports, Oxford, pp. 61–67.
•  Shigehara  N. and  Onodera  S. (1984) Skeletal remains of the domestic dogs from Tagara shell-mound, Miyagi Prefecture. Journal of the Anthropological Society of Nippon, 92: 187–210 (in Japanese with English abstract).
•  Shigehara  N.,  Eto  M. and  Baba  H. (1980) Human and dog remains from Ugasaki shell mound. In: Miyagi Prefecture Board of Education (ed.), Miyagi Prefecture Cultural Assets Survey Report 67: Kongoji Shell Mound, Ugasaki Shell Mound, Ugasaki No. 1 Tumulus. Miyagi Prefecture Board of Education, Sendai, pp. 167–182 (in Japanese).
•  Shigehara  N.,  Hiraguchi  T. and  Sakurai  H. (2004) Dogs and other small- and medium-sized mammals remains of the Mibiki site. In: Ishikawa Prefecture Buried Cultural Assets Center (ed.), Tatsuruhama Town Mibiki Site. Ishikawa Prefecture Buried Cultural Assets Center, Kanazawa, pp. 305–309 (in Japanese).
•  Shirai  K. (1992) Wild boar and methods of hunting. Shooter’s Japan, ’92–’93: 276–299 (in Japanese).
•  Ue  T. (1991) Japanese dogs preserved by the deep mountains. Anima, 221: 36–37 (in Japanese).
•  Watanabe  M. (1980) Remains of mammals and reptiles. In: Minamichita Town Board of Education (ed.), Mazukari Shell Mound. Minamichita Town Board of Education, Minamichita Town, pp. 83–84 (in Japanese).