Endocranial proportions and postorbital morphology of the Minatogawa I and IV Late Pleistocene Homo sapiens crania from Okinawa Island, Japan

Abstract

The purpose of this paper is to clarify the endocranial and postorbital morphologies of the terminal Pleistocene Minatogawa people, and obtain insights into their evolutionary background and genealogical relationships. The Minatogawa I and IV crania were compared with 83 Homo sapiens specimens (including 19 prehistoric Jomon). Metric comparative analyses and observational evaluations revealed that the two Minatogawa endocrania are characterized by a common suite of features (or tendency), including a small endocranial volume, relatively low endocranial shape, distinctly broad temporal region (qualitatively associated with the strong temporal bulge), weak parietal boss, and weakly swollen frontal bulge. Postorbital constriction was confirmed to be strong, relative to both upper facial breadth and maximum cranial breadth. Partial correlation analysis and bivariate comparisons were performed to examine the possible associations of Minatogawa’s strong postorbital constriction. The results suggest that constriction relative to the face is predominantly due to a large facial breadth (frontal endocranium not so narrow in Minatogawa I), and also in part to strong neuro-orbital disjunction, while constriction relative to the posterior cranium is largely attributable to endocranial shape. Comparisons with modern/recent H. sapiens materials and limited but informative outgroup specimens (Skhul V, and an example of Homo erectus, Daka) suggests that some of the features characteristic of Minatogawa are possibly ancestral retentions of early H. sapiens (e.g. strong temporal bulge, marked postorbital constriction relative to the face), while others are probably derived population features (e.g. a small endocranial volume, weakly swollen frontal bulge, marked postorbital constriction relative to the posterior cranium). In overall endocranial proportions, the Jomon tend to lie closer to Minatogawa than does the modern Japanese, but such morphologies were also found in individuals of other populations, and thus the similarity does not necessarily support the hypothesis of Minatogawa–Jomon genealogical closeness.

Introduction

The human fossils recovered from the Minatogawa limestone quarry fissure in Okinawa Island in 1970–1971 (Suzuki and Hanihara, 1982) are so far the best preserved pre-Jomon human remains of the Japanese archipelago. They include four partial skeletons (Minatogawa I–IV) and other fragmentary bones. Based on associated fauna and ¹⁴C dating (uncalibrated dates of 18250 ± 600 and 16600 ± 300 BP), the Minatogawa series have been considered to derive from the terminal Pleistocene (Suzuki and Hanihara, 1982). A recent re-evaluation of existing records confirmed that the dated charcoal fragments were collected from the clay unit that yielded Minatogawa individuals II, III, and IV (considered to be female), while individual I (considered to be male) and some other fragments were apparently recovered from slightly lower in the fissure sediments (Suwa et al., 2011a). Other chronological indicators include the wild boar and extinct deer remains recovered from the deposits that contained the human fossils. Elemental relative dating of these fossils (Matsu’ura and Kondo, 2011) and ¹⁴C-based chronology of wild boar remains at a newly discovered nearby fissure site (Yamasaki et al., 2010) considered together with the above-mentioned ¹⁴C dates suggest a most likely age of the Minatogawa human fossils of about 15000–20000 BP (or possibly extending slightly older) (also reviewed in Kaifu and Fujita, 2012).

The skeletal morphology of the Minatogawa human fossils was first described and evaluated in a monograph by Suzuki and Hanihara (1982). Aspects discussed in that monograph and thereafter include population origins, adaptation, presence or absence of primitive retentions, and genealogical relationship with the Jomon people (the Early–Middle Holocene inhabitants in the Japanese archipelago) (Suzuki and Hanihara, 1982; Baba and Narasaki, 1991; Yamaguchi, 1992; Baba et al., 1998; Brown, 1999; Baba, 2000a, b, 2002; Kaifu et al., 2011; Kubo et al., 2011; Mizoguchi, 2011; Saso et al., 2011; Suwa et al., 2011b). The earlier comparative studies suggested that the Minatogawa humans exhibit morphological similarities with the Late Pleistocene or Early Holocene inhabitants of southern China (Liujiang) or Southeast Asia (Wadjak). Similarities of the Minatogawa fossils with the Jomon-period prehistoric Japanese were also documented, thereby suggesting either a direct ancestor–descendent or close genealogical relationship between the two populations (Suzuki, 1982; Yamaguchi, 1992; Baba et al., 1998). However, in a recent detailed study of the mandible, Kaifu et al. (2011) pointed out that the Minatogawa series is distinct from the Jomon condition and resembles Australo-Melanesians in certain features such as a distinct alveolar prognathism and a small-sized mandibular ramus. They considered this as suggestive of a Southeast Asian origin of the Minatogawa people, perhaps from a population source with Australo-Melanesian affinities. They also suggested that this may indicate a lack of genealogical continuity from Minatogawa to the Jomon, especially with the increasing genetic, morphological, and archeological evidence that supports an East or Northeast Asian origin of the Jomon people (e.g. Adachi et al., 2009; Hanihara and Ishida, 2009; Nakashima et al., 2010; Kaifu and Fujita, 2012). Probable population hiatus in insular environments of the Okinawa islands was initially suggested by Takamiya (1996).

Further morphological comparisons and analyses potentially inform these issues. Saso et al. (2011) and Suwa et al. (2011b) examined some other Minatogawa morphologies (glabellonasal and mandibular tooth roots, respectively) distinct from the Jomon condition. On the other hand, it remains true that the Minatogawa series do share many aspects of craniofacial form with the Jomon people (Suzuki, 1982; Baba et al., 1998), rather than with Australo-Melanesians or early Homo sapiens so far known from Southeast Asia (Brown, 1999; Matsumura, 2006). Thus, pending further analyses and discoveries, the origin and genealogical relationships of the Minatogawa people remain unclear (see also Kaifu and Fujita, 2012).

In his initial studies, Suzuki (1982) estimated the endocranial volumes (ECVs) of Minatogawa I, II, and IV, by filling millet seed into the endocranial cavity, to be 1390, 1170, and 1090 cc, respectively. Recently, we were able to estimate more accurately the Minatogawa I and IV endocranial volumes from reconstructions based on serial microfocal X-ray computed tomography (micro-CT) scans (Kubo et al., 2008; Kubo and Kono, 2011). Revised ECV estimations of 1335 and 1170 cc were obtained for Minatogawa I and IV, respectively. Despite the upward revision in the case of Minatogawa IV, the newer ECV estimates confirm that the Minatogawa humans had smaller ECVs than the Jomon or the modern Japanese.

Using the same micro-CT-based digital data, we recently described the endocranial morphology of Minatogawa I and noted some distinct features such as a weak parietal boss and strong temporal bulge (Kubo et al., 2011). This was tentatively interpreted as a primitive condition of H. sapiens, possibly characteristic of early H. sapiens in general (Kubo et al., 2011). In the present study, we compare the endocranial shape and proportions of Minatogawa I and IV with those of Jomon, modern Japanese, Australians, and some other H. sapiens specimens in order to clarify the shared endocranial features of the two Minatoagwa crania, to evaluate the affinity of endocranial morphology between Minatogawa and Jomon, and thereby to obtain insights into the microevolutionary background and subsequent genealogical relationship of the Minatogawa people.

In addition, we investigate details of the often emphasized (e.g. Suzuki, 1982; Baba et al., 1998) postorbital morphology of the Minatogawa crania, exhibiting a marked postorbital constriction relative to upper facial breadth compared with the Jomon and modern Japanese (Suzuki, 1982). Baba et al. (1998) related the marked postorbital constriction of Minatogawa I to the extraordinary deep temporal fossa in its anterior portion. This and the temporal line placed high on the frontal bone indicate a strongly developed anterior portion of the temporal muscle. On the other hand, although they considered the Minatogawa I cranium to have a small frontal bone, they did not explicitly discuss possible relationships between postorbital constriction and brain size or shape.

Evolutionary changes of postorbital constriction in the human lineage have been considered related to that of masticatory apparatus size as well as encephalization (e.g. Weidenreich, 1941; Aiello and Dean, 1990). In addition, some researchers have mentioned a possible effect of ‘neuro-orbital disjunction’ (separation between orbit and frontal endocranial region); a larger separation potentially reduces postorbital breadth, leading to stronger constriction (Moss and Young, 1960; Cameron and Groves, 2004; Curnoe, 2009). For example, Curnoe (2009) attributed the postorbital constriction of robust Australian H. sapiens such as WLH50 to both the strong neuro-orbital disjunction and the large volume of masticatory muscles. However, the associations between postorbital breadth/constriction and brain size/shape or neuro-orbital disjunction have not previously been explicitly or quantitatively examined.

Another matter stems from the term ‘postorbital constriction’; in many cases (including the aforementioned studies) it denotes constriction relative to facial breadth (e.g. upper facial breadth, biorbital breadth), while sometimes it means the constriction relative to breadth at a more posterior part of the cranium (e.g. biparietal breadth). These two ‘types’ of postorbital constriction probably represent different aspects of cranial morphology. Previous studies have reported a general trend for postorbital constriction to reduce relative to the face: from the australopithecines to early Homo (Rak, 1983; Asfaw, 1987; Kimbel et al., 2004), from early Homo to early H. erectus (H. ergaster) (Cameron et al., 2004), and from archaic Homo to H. sapiens as well as within H. sapiens (Rightmire, 2008, 2009). On the other hand, the degree of postorbital constriction relative to the posterior cranium was reported to be relatively comparable between australopithecines and early Homo (Wood, 1991) and reduced within archaic Homo in Africa and Indonesia, respectively (Gilbert, 2008; Kaifu et al., 2008), and there seems no report on the H. sapiens condition. In the present study, we metrically examine whether and how postorbital breadth and the two measures of postorbital constriction are related to key endo- and ectocranial dimensions and proportions, including facial and endocranial breadths and degree of neuro-orbital disjunction. We then interpret the postorbital constriction of the Minatogawa crania in light of these results.

Materials and Methods

Materials and data preparation

The cranial CT data of Minatogawa I and IV and 82 comparative Jomon/modern H. sapiens specimens were obtained, and their cranial and endocranial polygon surface models were reconstructed. The Holocene sample includes 19 prehistoric Jomon (10 male and 9 female) and 63 modern (late 19th and early 20th centuries) crania: 52 Japanese (32 male and 20 female), 6 Australians (4 male and 2 female), and one male each of African, Buriat, European, Hawaiian, and Peruvian crania. The Jomon specimens derive from various districts of Honshu Island (Chugoku, Chubu, Kanto, and Tohoku) and chronologically span from Early to Final Jomon phases (7250 cal BP to 2360 cal BP: Kudo, 2004) (Table 1). The CT data of seven Jomon crania were taken from the database of the Laboratory of Physical Anthropology, Kyoto University (Makishima and Ogihara, 2009) and kindly provided to us for use by Dr. N. Ogihara. These were scanned by a helical CT scanner (XVision TSX-002/4I, Toshiba Medical Systems Co., Tokyo) at the Kyoto University laboratory, and reconstructed in matrices of isotropic 0.5 mm sized voxels. The other specimens were housed in the University of Tokyo, or the National Museum of Nature and Science, Tokyo, and were scanned by a microfocal X-ray CT system (TXS225-ACTIS, TESCO Co., Tokyo) at the University Museum, the University of Tokyo. The CT data of the Minatogawa I and IV crania were reconstructed in isotropic matrices of 0.380 and 0.350 mm sized voxels, respectively, while isotropic voxel datasets 0.350–0.450 mm in size were acquired for the comparative sample (see Kubo and Kono, 2011, and Kubo et al., 2011 for details of the parameters used in CT-scanning and reconstruction). In addition to these, the micro-CT data of Daka, a H. erectus cranium (Gilbert et al., 2008), and the CT data of Skhul V, an early Late Pleistocene H. sapiens cranium (available at http://www.peabody.harvard.edu/node/446), were also consulted as ancestral examples relative to later H. sapiens. Despite the limited scope that Skhul V and Daka provide regarding either early H. sapiens or H. erectus conditions, we neverthless find their inclusion insightful and therefore necessary in providing clues as to evolutionary morphoclines that apply to H. sapiens cranial form.

Table 1 Comparative Jomon crania used in this study

Site	n		Phase	Locality	Collectiona
	M	F
Sano	1	1	Early Jomonb	Chiba	UMUT
Hikozaki	2	1	Early Jomonc	Okayama	NMNS
Wakaumi	1		Middle Jomond	Ibaraki	NMNS
Yosekura	1		Late Jomone	Hiroshima	UMUT
Hosoura	1		Middle–Final Jomonf	Iwate	UMUT
Ikawazu	1		Late–Final Jomong	Aichi	NMNS
Miyano	1		Final Jomonh	Iwate	NMNS
Hobi		2	Final Jomoni	Aichi	UMUT
Tsukumo	2	5	Final Jomonj	Okayama	KU

a  UMUT, University Museum, University of Tokyo; NMNS, National Museum of Nature and Science, Tokyo; KU, Laboratory of Physical Anthropology, Kyoto University.

b  Yoneda et al. (2005).

c  Yamasaki and Takahashi (2007).

d  Kajigayama and Baba (1999).

e  Fukase et al. (2008).

f  Iwate Prefecture (1961).

g  Editorial Committee of the History of Aichi Prefecture (2002).

h  Hayashi et al. (1981).

i  Suwa et al. (2003).

j  Ikeda (1991).

For surface model extraction, the endocranial region was first segmented by half maximum height thresholding between the CT values of air and endocranial bone, and subsequently by slice-by-slice manual adjustment of bone–air boundaries using the commercially available softwares Analyze 8.1 (Mayo Clinic, MN) and Amira 5.2 (Visage Imaging, Inc., CA). After segmentation, the surface model was extracted by the marching-cube routine in Analyze 8.1. The Minatogawa IV cranium exhibits some postmortem distortion and much of the cranial base is missing. We thus reconstructed the endocranial model after correcting distortion and breakage by combining digital and manual restoration procedures (for details, see Kubo and Kono, 2011). Cranial surface models were also extracted by the same routine described above, but the resolution of the volume data was reduced to 0.600 mm for ease of data manipulation.

Although the Minatogawa I endocast has been described in some detail (Kubo et al., 2011), this is not the case for Minatogawa IV. We therefore show the basic orthogonal views and summarize the main morphological features of the Minatogawa IV endocast in Figure 1. These observations, when relevant, are integrated in the discussions of the present paper.

Figure 1

Orthogonal views of the Minatogawa IV endocast. The missing parts are dark colored. The length of the scale bar is 10 cm. The main features of the Minatogawa IV endocast are as follows: frontal petalia not clear; strong left occipital petalia; a weakly swollen frontal bulge leading to a laterally receding endocranial anterior outline in vertical view (resembling Minatogawa I); lack of marked parasagittal ridges (strong in Minatogawa I, Kubo et al., 2011); inferior protrusion of the left frontal cap (partly preserved, and similar to Minatogawa I); steep slope from vertex towards the occipital pole in lateral view (derived H. sapiens condition); relatively weak endocranial parietal boss (as in Minatogawa I); strong temporal bulge (as in Minatogawa I, but less extreme); lack of marked bilateral depression at anterior parietal region (contra Minatogawa I); cerebellar bulge lower and longer than in Minatogawa I; impression of the internal occipital crest narrow and deep (derived H. sapiens condition); horizontal sulcus discernible on the right cerebellar surface; obelic rami of middle meningeal artery from the anterior main branches; superior sagittal sinus runs into the right transverse sinus; and lack of occipital sinus impressions.

Assessment of endocranial morphology

Endocranial measurements (Table 2, Figure 2) were taken on the digital model, and recorded to the nearest 0.1 mm. The orientation was defined as follows: the midsagittal plane was determined solely by three landmarks (foramen caecum, internal lambda, and opisthion), but in some cases, minor adjustment was made by consulting other midsagittal and bilateral structures. The horizontal plane was then determined so that it is orthogonal to the midsagittal plane and passes through the foramen caecum and another point that corresponds to the base of the posterior cerebrum (see Kubo et al., 2011 for details).

Table 2 Metrics used in this study

Metrics			Definition and description
Length	§		Maximum length of the endocast in anteroposterior direction.
Width	§		Maximum width of the endocast. Parietal width or temporal width, whichever is greater.
Parietal width			Maximum width of the endocast at the dorsal part to the impression of the Sylvian crest.
Temporal width			Maximum width of the endocast at the ventral part to the impression of the Sylvian crest.
Temporal bossing	§		Temporal width–parietal width. An indicator of lateral protrusion of the endocranial temporal region relative to the endocranial parietal region.
Height	§		Vertical length from the bottom of the posterior cranial fossa to the top of the endocast, measured perpendicular to the horizontal plane†.
Frontal width	§	¶	Maximum width of the endocranial coronal-section at 20% of the length from the anteriormost point to the posteriormost point of the endocast.
Occipital width	§		Maximum width of the endocranial coronal-section at 80% of the length from the anteriormost point to the posteriormost point of the endocast.
Superior cerebral height			Vertical length from the horizontal plane† to the top of the endocast.
Frontal height	§		Vertical length from the horizontal plane† to the highest point on the endocranial coronal-section at 20% of the length from the anteriormost point to the posteriormost point of the endocast.
Occipital height	§		Vertical length from the horizontal plane† to the highest point on the endocranial coronal-section at 80% of the length from the anteriormost point to the posteriormost point of the endocast.
Occipital protrusion	§		Sagittal length from the posteriormost point of the cerebellar surface to the occipital pole.
PCF length	§		Sagittal length from the posterior edge of the opening of internal acoustic meatus to the posteriormost point of the cerebellar surface.
PCF width	§		Width of the posterior cranial fossa excluding the sinus.
PCF height	§		Vertical length from the horizontal plane† to the bottom of the posterior cranial fossa.
Endocranial volume			Volume of the endocranial cavity, shortened to “ECV”.
Width–length index			Width × 100/Length.
Height–length index			Height × 100/Length.
Height–width index			Height × 100/Width.
Frontal width index		¶	Frontal width × 100/Temporal width.
Occipital width index			Occipital width × 100/Temporal width.
Maximum cranial breadth		¶	Martin No. 8 (eu-eu).
Postorbital breadth		¶	Martin No. 9(1).
Upper facial breadth		¶	Martin No. 43 (fmt-fmt).
Pob-Ufb index		¶	Martin No. 9(1) × 100/Martin No. 43. An indicator of postorbital constriction relative to the upper face.
Pob-Xcb index		¶	Martin No. 9(1) × 100/Martin No. 8. An indicator of postorbital constriction relative to the posterior part of neurocranium.
Ufb-Fw index		¶	Martin No. 43 × 100/Frontal width. An indicator of relative width of the upper face to the endocranial frontal region.
Supraorbital protrusion		¶	Sagittal length in Frankfurt horizontal orientation from the anteriormost point of the endocranial cavity to that of the supraorbital ridge.
Orbital protrusion		¶	Sagittal length in Frankfurt horizontal orientation from the anteriormost point of the endocranial cavity to that of the superior margin of the orbital opening.

§  Metrics used in principal component analysis (PCA) (see text).

¶  Metrics used in assesment of postorbital morphology (see text).

†  The horizontal plane was determinded so that it became orthogonal to the midsagittal plane and passed through the foramen caecum and another point that correponds to the base of the posterior cerebrum, following Kubo et al. (2011).

Figure 2

Diagrammatic representation of measurements. The definitions are shown in Table 2.

We first metrically described the endocranial shape of Minatoagwa I and IV with reference to the Jomon and modern Japanese samples. Two-way ANOVA tests for the effects of sex and population group (Jomon vs. modern Japanese) were performed on each of raw measurements, the indices, and the size-standardized values (calculated by dividing the raw measurements by the cube root of ECV) of the Jomon and modern Japanese samples to investigate the effects of size-standardization on evaluations of population and sex differences. Since the results indicated that sex was insignificant in comparisons of the size-standardized variables (see below), we then proceeded to evaluate the two Minatogawa crania against the combined-sex comparative samples. This has the advantage of enabling better appreciation of variation in both the Jomon and Japanese samples and Minatogawa (albeit n = 2). This procedure also enables inclusion of fossils with unreliable sex attribution. Since a metric ECV did not fulfill the assumption of homogeneity of variance (Bartlett test, P < 0.05), the sex and interpopulation differences were also evaluated using the Mann–Whitney U-test. We then compared Minatogawa I and IV with the sex-combined comparative samples using the size-standardized values and some ratios. The shape characteristics of the Minatogawa and modern Japanese crania were evaluated by deviations (z-scores) from the Jomon mean.

Furthermore, principal component analysis (PCA) of 12 selected metrics (denoted by § in Table 2) was performed on the combined Minatogawa and comparative H. sapiens sample (n = 85) to summarize the variation patterns of endocranial shape in a comparative sample that represents broader geographical and deeper chronological (Skhul V) contexts than possible only with the Japanese samples. The principal components (PCs) were calculated on the correlation matrix of the size-standardized metrics. Only the PCs that explained over 10% of variance were interpreted by referring to significant factor loadings (P < 0.01, based on significance test for correlation between metrics and PC scores).

Assessment of postorbital morphology

The metrics used in the following analyses are denoted in Table 2. These include two measures of postorbital constriction (Pob-Ufb and Pob-Xcb indices) and the putative related variables. Ectocranial breadths were taken on the digital cranial model, while orbital/supraorbital protrusions (Figure 2), indicators of neuro-orbial disjunction, were taken on cranial and endocranial models, precisely superimposed with each other and oriented in the Frankfurt horizontal.

We first calculated partial correlation coefficients based on the combined Jomon/modern H. sapiens sample (n = 82) to examine whether and how postorbital breadth and constriction are affected by each of the following dimensions or proportions: endocranial frontal width, endocranial width proportion (frontal width index), upper facial breadth, and supraorbital/orbital protrusions. Variables controlled in partial correlations were selected a priori from putative relationships among these factors and postorbital breadth or constriction. For example, considering that the degree of anterior protrusion of the face relative to the endocranium is dependent on the size of the face (e.g. Lieberman, 2000), upper facial breadth (a measure representing size of the face) is controlled when examining effects of neuro-orbital disjunction (represented by orbital or supraorbital protrusions).

We then metrically compared the Minatogawa conditions with the Jomon/modern H. sapiens specimens as well as with Skhul V and an example of H. erectus (Daka). Two-way ANOVA tests for the effects of sex and population group on the measurements of the Jomon and modern Japanese samples were performed to clarify sex-independent shared Jomon tendencies with Minatogawa relative to the modern Japanese. Since two metrics (maximum cranial breadth and the Pob-Xcb index) did not fulfill the assumption of homogeneity of variance (Bartlett test, P < 0.05), their sex and interpopulation differences were also evaluated using the Mann–Whitney U-test. We also conducted bivariate analysis of key ecto- and endocranial dimensions/proportions that relate with postorbital constriction, in which we calculated the reduced major axis (RMA) based on the Jomon/modern H. sapiens comparative sample (n = 82) to evaluate whether and how the fossil specimens (Minatogawa, Skhul, and Daka) deviate from the variation of the later H. sapiens sample. All the above calculations were conducted using the statistical environment R (R Development Core Team, 2008) and Excel (Microsoft Co., WA).

Results

Endocranial dimensions and proportions

Table 3 shows the endocast measurements, indices, and ECV of Minatogawa I and IV as well as the comparative Jomon and modern Japanese samples, together with the deviations of the Minatogawa series relative to the Jomon sample. Table 4 summarizes the results of two-way ANOVA tests for the effects of sex and population group (Jomon and modern Japanese) on these endocranial metrics, together with results of the Mann–Whitney U-test for ECV (in the table footnotes). This summary shows that: (1) when size-standardized by the cube root of ECV, the sex differences are not significant in any variables we examined; (2) endocranial widths (especially temporal width and width of the posterior cranial fossa (PCF width), P < 0.001) are absolutely larger in Jomon than in modern Japanese, resulting in a relatively short, wide, and low endocranial proportion of the former (width–length and height–width indices, P < 0.05). In addition, in absolute and relative terms, temporal bossing and PCF length were larger (P < 0.01) while frontal height was lower (P < 0.001) in Jomon than in modern Japanese. The interaction effects of sex and population group (Jomon or modern Japanese) were not significant in any variables.

Table 3 Endocranial measurements of the Minatogawa, Jomon and modern Japanese crania

Measurement (linear, mm; ECV, cc)	Minatogawa I			Jomon male (n = 10)		Japanese male (n = 32)		Minatogawa IV			Jomon female (n = 9)		Japanese female (n = 20)
		Z relative to Jomon							Z relative to Jomon
		Raw §	Std. ¶	Mean	SD	Mean	SD		Raw †	Std. ¶	Mean	SD	Mean	SD
Length	163.6	−1.7	−0.1	169.0	3.3	169.6	5.3	161.3	−0.7	+1.5	164.4	4.4	161.0	5.1
(left)	164.5			169.5	3.6	170.3	5.7	163.4			164.6	4.6	162.1	5.3
(right)	162.6			168.5	3.1	169.0	5.2	159.2			164.3	4.2	159.9	5.0
Width	140.4	+0.7	+2.5	137.8	3.6	132.7	4.7	129.8	−0.8	+1.0	132.5	3.3	128.1	4.1
Parietal width	129.3	−0.6	+0.3	132.8	5.4	129.8	4.2	122.3	−2.4	−0.1	128.3	2.5	125.3	4.9
Temporal width	140.4	+0.7	+2.5	137.8	3.6	132.6	4.8	129.8	−0.8	+1.0	132.5	3.3	128.1	4.0
Temporal bossing	11.1	+1.7	+2.1	5.0	3.5	2.8	2.0	7.5	+1.1	+1.1	4.2	3.0	2.8	2.3
Height	122.8	−1.8	−0.5	129.3	3.7	128.8	4.2	116.7	−2.1	−0.8	123.8	3.5	122.3	4.3
Frontal width	103.2	−1.5	+0.3	107.5	2.9	103.8	3.2	(97.9)	−1.1	−0.1	101.0	2.9	99.6	2.6
Occipital width	107.7	−0.9	+0.5	110.8	3.3	107.1	4.5	97.2	−3.2	−2.0	105.8	2.7	102.6	4.0
Superior cerebral height	86.8	−3.8	−1.2	95.1	2.2	95.6	4.2	87.1	−0.6	+0.1	89.7	4.6	90.0	3.8
Frontal height	68.7	−2.0	−0.4	72.9	2.1	76.7	3.4	68.3	−0.2	+0.7	68.9	2.9	72.8	3.4
Occipital height	75.9	−2.0	−0.8	81.7	2.8	82.3	4.5	71.1	−1.6	−1.2	77.6	4.1	77.4	3.8
Occipital protrusion	14.5	−0.3	−0.1	15.2	2.7	13.5	2.8	13.8	−0.4	−0.2	14.8	2.7	13.7	3.0
(left)	15.1			16.1	2.8	14.3	3.4	17.0			15.1	2.9	14.9	3.2
(right)	13.9			14.2	2.9	12.7	2.8	10.5			14.5	2.9	12.5	3.1
PCF length	54.4	−2.6	−2.8	61.6	2.7	60.0	2.6	57.2	−1.8	−0.0	60.3	1.7	56.9	2.1
PCF width	101.2	−2.6	−0.8	108.5	2.8	103.2	4.6	(101)	−0.5	+0.5	103.3	4.5	96.9	3.1
PCF height	36.0	+0.5	+0.8	34.2	3.8	33.2	3.3	29.6	−1.5	−0.8	34.1	2.9	32.3	3.4
Endocranial volume	1335	−2.5		1495.0	64.3	1475.3	102.3	1170	−3.4		1326.9	46.8	1273.5	65.4
Width–length index	85.8	+1.4	+1.8	81.6	3.0	78.3	3.8	80.5	+0.0	−0.2	80.6	2.4	79.7	3.8
Height–length index	75.1	−0.6	−0.3	76.5	2.3	76.0	2.8	72.3	−0.9	−1.3	75.4	3.3	76.0	2.7
Height–width index	87.5	−1.7	−1.6	93.9	3.9	97.2	4.0	89.9	−0.9	−1.0	93.5	4.2	95.5	4.7
Frontal width index	73.5	−3.2	−2.1	78.1	1.4	78.3	2.7	(75.4)	−0.5	−1.0	76.3	1.6	77.8	2.1
Occipital width index	76.7	−1.5	−1.5	80.4	2.5	80.8	2.4	74.9	−2.1	−2.2	79.9	2.4	80.1	2.2

§  Deviations (Z-scores) relative to the Jomon male sample means were calculated based on raw measurements.

†  Deviations relative to the Jomon female sample means were calculated based on raw measurements.

¶  Deviations relative to the sex-combined Jomon sample means. In the metrics other than five indices, they were calculated based on the size-standardized values for ECV, while the indices were based on raw values.

Parenthetic figures represent the estimated values.

Table 4 Two-way ANOVA summary for endocranial measurements.

	Differences between males (M) and females (F)			Differences between Jomon (Jo) and modern Japanese (mJ)
	Raw		Size-standardized	Raw		Size-standardized
Length	M > F	***	ns		ns		ns
Width	M > F	***	ns	Jo > mJ	***	Jo > mJ	***
Parietal width	M > F	***	ns	Jo > mJ	*		ns
Temporal width	M > F	***	ns	Jo > mJ	***	Jo > mJ	***
Temporal bossing		ns	ns	Jo > mJ	**	Jo > mJ	**
Height	M > F	***	ns		ns		ns
Frontal width	M > F	***	ns	Jo > mJ	*	Jo > mJ	*
Occipital width	M > F	***	ns	Jo > mJ	*	Jo > mJ	**
Superior cerebral height	M > F	***	ns		ns		ns
Frontal height	M > F	***	ns	mJ > Jo	***	mJ > Jo	***
Occipital height	M > F	***	ns		ns		ns
Occipital protrusion		ns	ns		ns		ns
PCF length	M > F	***	ns	Jo > mJ	***	Jo > mJ	**
PCF width	M > F	***	ns	Jo > mJ	***	Jo > mJ	***
PCF height		ns	ns		ns		ns
Width–length index		ns		Jo > mJ	*
Height–length index		ns			ns
Height–width index		ns		mJ > Jo	*
Frontal width index		ns			ns
Occipital width index		ns			ns
Endocranial volumea	M > F	***			ns

Significant level of two-way ANOVA were denoted by: ns, not significant;

*  P < 0.05;

**  P < 0.01;

***  P < 0.001.

Interaction effects of sex (M, F) and population group (Jo, mJ) were not significant in any variables.

a  An assumption of ANOVA (homogeneity of variance) was statistically rejected (Bartlett test, P < 0.05). Mann–Whitney U-test showed significant differences between male and female (P < 0.001, in both Jomon and modern Japanese) and population groups (Jo > mJ, P < 0.05, only in female).

Due to the great difference in ECV, most of the endocranial dimensions of Minatogawa I (IV) are absolutely smaller than the Jomon male (female) mean (Table 3). When the endocranial shape was compared by size-standardized measurements and indices (Figure 3), we found that, relative to the Jomon and mostly contrary to the modern Japanese condition, the two Minatogawa endocrania were characterized by distinctly large maximum (or temporal) width and temporal bossing as well as small height–width, frontal and occipital width indices, and a tendency for low (height and occipital height) endocranial proportions. In addition, some other metrics (frontal, parietal, PCF widths and frontal height) of Minatogawa are comparable to the Jomon rather than the modern Japanese. On the other hand, the two Minatogawa endocasts are different from each other in some aspects: the Minatogawa I endocast is wide in its overall proportion (size-standardized width and width–length index are 2.5 standard deviation (SD) and 1.8 SD units above the Jomon male means, respectively), while the Minatogawa IV endocast is long (size-standardized length is 1.5 SD above the Jomon female mean); the above-noted large temporal bossing is more distinct in Minatogawa I (2.1 SD above the Jomon male mean) than in Minatogawa IV; and the Minatogawa I has a very short PCF region (size-standardized PCF length is 2.8 SD below the Jomon male mean), while the Minatogawa IV does not share this condition.

Figure 3

Deviation plots of endocranial metrics relative to the sex-combined Jomon mean, based on size-standardized values. Filled square, filled circle, and open circle represent Minatogawa I, Minatogawa IV, and the sex-combined modern Japanese mean, respectively.

Figure 4 and Table 5 show the results of the PCA of endocranial metrics, based on the correlation matrix of the size-standardized values. The first three principal components (PCs) account for 29.5%, 17.6%, and 13.7% of the variance, respectively. A small PC1 score is related with a short, wide and low endocranial proportion with a strong temporal bulge. The Jomon tended to have relatively small PC1 scores distinct when compared with the modern Japanese or the modern Australians, which have large PC1 scores. The Minatogawa I and IV, Skhul V and the single examples of Buriat and European individuals had smaller PC1 scores than the Jomon mean, although within the range of variation of the Jomon. A small PC2 score is related with a long and low endocranial proportion with narrow occipital region, strong temporal bulge, and strong occipital protrusion. The PC2 score of Minatogawa I was comparable with the Jomon and modern Japanese means, while the modern Australians tended to have smaller PC2 scores. Minatogawa IV, Skhul V, and an example of Buriat had smaller PC2 scores than the modern Australian mean, although these were within the range of variation of the small Australian sample. In the PC3 scores, Minatogawa I and IV, Skhul V, and the modern Japanese and Australian means all lay relatively close to the Jomon mean within the range of variation of the Jomon.

Figure 4

Summary plots of principal component analysis based on the correlation matrix of 12 size-standardized endocranial measurements. (A) PC1 score vs. PC2 score. (B) PC1 score vs. PC3 score. The group means of the Jomon, modern Japanese, and modern Austaralian are represented by larger symbols, and their ranges of variation are represented by convex hulls.

Table 5 Factor loadings of the PCA based on size-satandardized endocranial measurements

	PC1	PC2	PC3
Length	0.30	−0.73	0.47
Width	−0.81		−0.35
Temporal bossing	−0.41	−0.41
Height	0.49	0.52	0.40
Frontal width	−0.68
Occipital width	−0.63	0.40	−0.32
Frontal height	0.66		−0.37
Occipital height	0.69		−0.41
Occipital protrusion	0.36	−0.57
PCF length	−0.37	−0.34	0.50
PCF width	−0.54
PCF height	−0.29	0.66	0.60
Contribution (%)	29.5	17.6	13.7

Only significant factor loadings (P < 0.01, based on significance test for correlation between measuments and PC scores) are shown.

Postorbital constriction and related morphologies

Table 6 shows the correlation and partial correlation coefficients between postorbital breadth and the two indices of constriction on the one hand and related variables on the other, using the combined Jomon and modern H. sapiens sample. The detected significant partial correlations denote that: (1) a large endocranial frontal width is associated with a wide postorbital breadth (15% of the variation is accounted for when control variables were held constant), large Pob-Ufb index (15%), and small Pob-Xcb index (26%); (2) a large endocranial frontal width index is associated with a large Pob-Xcb index (33%); (3) a large upper facial breadth is associated with a wide postorbital breadth (56%), small Pob-Ufb index (7%), and large Pob-Xcb index (23%); (4) supraorbital protrusion (controlled for upper facial breadth and frontal width) does not show significant partial correlations with any of the postorbital breadth and constriction measures; and (5) a large orbital protrusion is associated with a narrow postorbital breadth (5%) and small Pob-Ufb index (5%), indicating a slight but significant effect of neuro-orbital disjunction on postorbital constriction.

Table 6 Correlation/partial correlation coefficients related to postorbital breadth/constriction based on a mixed Jomon/modern H. sapiens sample (n = 82)

possible related variables	Postorbital breadth/constriction			Control variables
	Postorbital breadth	Pob-Ufb index	Pob-Xcb index
	P	P	P
Correlations
Frontal width (endocranial)	0.551 ***	0.373 ***	−0.190
Frontal width index (endocranial)	0.236 *	0.030	0.603 ***
Upper facial breadth	0.771 ***	−0.182	0.530 ***
Supraorbital protrusion	0.240 *	−0.347 **	0.297 **
Orbital protrusion	−0.005	−0.389 ***	0.158
Partial correlations
Frontal width (endocranial)	0.387 ***	0.391 ***	−0.505 ***	Upper facial breadth, orbital protrusion
Frontal width index (endocranial)	0.095	0.097	0.575 ***	Upper facial breadth, orbital protrusion
Upper facial breadth	0.747 ***	−0.263 *	0.482 ***	Frontal width (endocranial), orbital protrusion Frontal width index (endocranial), orbital protrusion
Supraorbital protrusion	−0.155	−0.157	0.057	Frontal width (endocranial), upper facial breadth Frontal width index (endocranial), upper facial breadth
Orbital protrusion	−0.217 *	−0.219 *	−0.006	Frontal width (endocranial), upper facial breadth Frontal width index (endocranial), upper facial breadth

Statistical significance:

*  P < 0.05,

**  P < 0.01,

***  P < 0.001.

Table 7 shows the related endo- and ectocranial measurements of the Minatogawa and the comparative Homo samples, together with the deviations of the Minatogawa from the Jomon mean. Relative to the Jomon, the two Minatogawa crania were characterized by absolutely small endocranial frontal width and postorbital breadth, marked postorbital constrictions represented by Pob-Ufb and Pob-Xcb indices, and large orbital and supraorbital protrusions. In particular, deviations of endocranial frontal width index, Pob-Ufb index, and orbital protrusion are more distinct in Minatogawa I (3.2 SD below, 2.7 SD below, and 2.9 SD above the Jomon male means, respectively) than in Minatogawa IV. In addition, Minatogawa I has a very wide face compared with the Jomon (upper facial breadth and Ufb-Fw index are 2.3 SD and 2.1 SD above the Jomon male means, respectively). Table 8 summarizes the results of two-way ANOVA tests for the effects of sex and population group (Jomon and modern Japanese) on these metrics, together with results of the Mann–Whitney U-test for maximum cranial breadth and Pob-Xcb index (in the table footnotes). Sex differences were not detected in any of the ratio metrics including the two measures of postorbital constriction or supraorbital protrusion, while orbital protrusion was larger in females (P < 0.05). All breadth measurements were significantly larger in males (P < 0.01 or P < 0.001). Relative to the modern Japanese, the Jomon exhibits a slight tendency towards the Minatogawa conditions in a small Pob-Xcb index (P < 0.05 only in female, Mann–Whitney U) and large supraorbital protrusion (P < 0.001, two-way ANOVA).

Table 7 Postorbital constriction and the possible related variables of the Minatogawa crania and the comparative samples

Linear measurement (mm), and index	Minatogawa I		Jomon male (n = 10)		Japanese male (n = 32)		Australian male (n = 4) Mean	Minatogawa IV		Jomon female (n = 9)		Japanese female (n = 20)		Australian female (n = 2) Mean	Skhul V	Daka (H. erectus)
		Z §	Mean	SD	Mean	SD			Z †	Mean	SD	Mean	SD
Frontal width (endocranial)	103.2	−1.5	107.5	2.9	103.8	3.2	97.4	(97.9)	−1.1	101.0	2.9	99.6	2.6	92.9	105.4	96
Frontal width index (endocranial)	73.5	−3.2	78.1	1.4	78.3	2.7	79.6	(75.4)	−0.5	76.3	1.6	77.8	2.1	80.0	77.0	(77.9)
Maximum cranial breadth	147.6	+0.5	146.2	2.8	139.2	4.7	128.4	138.0	−1.6	141.2	2.1	135.8	4.3	122.4	143.6	133
Postorbital breadth	92.9	−1.0	97.4	4.7	94.5	4.0	97.5	(88.4)	−1.4	92.5	2.8	91.9	3.8	90.2	103.7	95
Upper facial breadth	112.1	+2.3	106.6	2.4	103.1	3.9	110.5	102.0	−0.4	103.1	2.8	100.2	2.8	103.3	(124)	124
Pob-Ufb index	82.9	−2.7	91.3	3.1	91.7	2.6	88.3	(86.7)	−1.9	89.7	1.6	91.7	2.3	87.3	(83.6)	76.6
Pob-Xcb index	62.9	−1.0	66.7	4.0	67.9	3.3	76.0	(64.1)	−1.1	65.5	1.2	67.7	2.8	73.8	72.2	71.4
Ufb-Fw index	108.6	+2.1	99.3	4.4	99.4	3.8	113.5	(104.2)	+0.5	102.2	4.6	100.7	2.7	111.2	(117.6)	129.2
Supraorbital protrusion	12.9	+1.3	9.0	2.9	7.3	1.7	12.4	11.4	+1.3	9.1	1.8	6.5	2.0	13.2	(16.6)	22.2
Orbital protrusion	7.5	+2.9	1.2	2.2	0.3	2.0	3.6	3.0	+0.6	2.2	1.4	1.3	2.2	5.5	8	14

§  Deviations relative to the Jomon male sample means were calculated based on raw measurements.

†  Deviations relative to the Jomon female sample means were calculated based on raw measurements.

Parenthetic figures represent the estimated values.

Maximum cranial breadth, postorbital breadth, and upper facial breadth of Daka are from Asfaw et al. (2008).

Table 8 Two-way ANOVA summary for the metrics of postorbital constriction and possible related features

	Differences between males (M) and females (F)		Differences between Jomon (Jo) and modern Japanese (mJ)
Frontal width (endocranial)	M > F	***	Jo > mJ	*
Frontal width index (endocranial)		ns		ns
Maximum cranial breadtha	M > F	***	Jo > mJ	***
Postorbital breadth	M > F	**		ns
Upper facial breadth	M > F	***	Jo > mJ	***
Pob-Ufb index		ns		ns
Pob-Xcb indexa		ns	mJ > Jo	*
Ufb-Fw index		ns		ns
Supraorbital protrusion		ns	Jo > mJ	***
Orbital protrusion	F > M	*		ns

Significant level of two-way ANOVA were denoted by: ns, not significant;

*  P < 0.05;

**  P < 0.01;

***  P < 0.001.

Interaction effects of sex (M, F) and population group (Jo, mJ) were not significant in any variables.

a  An assumption of ANOVA (homogeneity of variance) was statistically rejected (Bartlett test, P < 0.05). Mann–Whitney U-test showed significant differences between male and female (P < 0.01 in Jomon; P < 0.05 in modern Japanese) and population groups (Jo > mJ, P < 0.001 in male; P < 0.01 in female) in maximum cranial breadth as well as a significant difference in Pob-Xcb index (mJ > Jo, P < 0.05 only in female).

Figure 5A is a bivariate plot of the Pob-Ufb and Pob-Xcb indices. The Daka specimen, an example of H. erectus, exhibits a distinctly marked postorbital constriction represented by the Pob-Ufb index, followed by Skhul V, the Minatoagwa series, and some modern Australians, the expression of whose postorbital constriction is marked relative to the Jomon and modern Japanese. On the other hand, pos-torbital constriction represented by the Pob-Xcb index was found to be even weaker in Daka, Skhul V, and modern Australians than in the Jomon and modern Japanese means. Figure 5B shows a bivariate plot of supraorbital and orbital protrusions. Daka exhibited distinctly large orbital and supraorbital protrusions (strong neuro-orbital disjunction), followed by Skhul V, the Minatogawa series, and some Australian crania, the degree of whose orbital/supraorbital protrusion is marked relative to the Jomon and modern Japanese. Endocranial frontal width, as noted above (Table 6), accounts for a significant part of the variation of the Pob-Ufb index among the Jomon/modern H. sapiens specimens, and a small Pob-Ufb index of Minatogawa IV seems sufficiently explained by its narrow endocranial frontal width (Figure 5C). However, the Pob-Ufb index of Minatogawa I, Skhul V, and Daka was considerably smaller than expected from their endocranial frontal widths. These early H. sapiens and H. erectus crania had considerably larger upper facial breadths than expected from their postorbital breadths (Figure 5D), and the marked postorbital constriction (represented by the Pob-Ufb index) is explained by their large facial breadth relative to endocranial frontal width (Ufb-Fw index) (Figure 5E). The endocranial frontal width index accounts for the variation of the ectocranial Pob-Xcb index (Table 6), and the Pob-Xcb indices of Minatogawa I and IV are consistent with the values expected from the RMA derived from the Jomon/modern H. sapiens specimens (Figure 5F). Skhul V, Daka, and the modern Australian specimens tend to have larger Pob-Xcb index values than expected from the RMA on endocranial frontal width index (Figure 5F). Within H. sapiens, this may in part be accounted for by their relatively large upper facial breadth (that is associated with postorbital breadth in our modern H. sapiens sample) (Figure 5G).

Figure 5

Bivariate plots of selected endocranial and ectocranial metrics. Reduced major axes and correlation coefficients with the significance based on 82 Jomon/modern H. sapiens specimens are shown. Vertical axis of (G) is residual of RMA regression of Pob-Xcb index on endocranial frontal width index (F).

Discussion

Endocranial morphology of the Minatogawa specimens

The present metric comparisons showed that, compared with the Jomon and modern Japanese conditions, the Minatogawa I and IV endocasts share some characteristic features, including a small ECV, a relatively low endocranial shape, and a distinctly broad temporal region relative to the frontal/parietal/occipital region (qualitatively associated with the strong temporal bulge). In addition, our qualitative observations (Figure 1, and Kubo et al., 2011) found other features that characterize the two Minatogawa endocasts relative to the Jomon and modern Japanese conditions; in particular, a weak endocranial parietal boss, weakly swollen lateral frontal bulge, and a strongly inferiorly protruding frontal cap. Among these, a relatively low endocranial shape, strong temporal bulge, and a weak parietal boss are possible ancestral retentions of H. sapiens in general, because these are shared with earlier Late Pleistocene H. sapiens examples such as Skhul V and Qafzeh 9 (Kubo et al., 2011; and the present study), while the others (a small ECV as well as the frontal bulge and cap morphologies) are probably derived. We note that this is a tentative interpretation and needs to be confirmed and/or refined by further comparative studies based on broader samples in geographical and chronological terms.

Interpretation of postorbital constriction from morphological associations

The present results are based on a small comparative fossil sample and can only provide limited insights into the evolutionary changes of postorbital constriction from archaic Homo to H. sapiens and/or within H. sapiens. However, the following findings are noteworthy: (1) the possible contribution of neuro-orbital disjunction to postorbital breadth and constriction (metrically detected in mixed population H. sapiens sample); (2) the two measures of postorbital constriction, one relative to the face and the other relative to the posterior cranium, are differently related to ecto- and endocranial proportions; (3) the Minatogawa I and IV crania exhibit marked expression of both measures of postorbital constriction, compared with most of the Jomon and modern H. sapiens specimens; and (4) postorbital constriction relative to the posterior cranium is not necessarily marked in H. erectus (Daka) or early H. sapiens (Skhul V).

Our results show that the frontal endocranium of Minatogawa I is not so narrow, and does not account for their strong postorbital constriction relative to the face (as is also the case in Skhul and Daka). The marked expression of this measure of constriction in Minatogawa I is better considered due to a wide (and protruding) upper face rather than a narrow frontal endocranium, the latter previously implied from the small/narrow frontal squama (cf. Baba et al., 1998). Furthermore, taking into account the results of the partial correlation analysis, we suggest that the narrow postorbital breadth and marked postorbital constriction relative to upper facial breadth of the Minatogawa (and Australian) crania are also in part due to a large neuro-orbital disjunction.

Asfaw et al. (2008) reported that postorbital constriction relative to maximum cranial breadth is moderate in Daka compared with other H. erectus and archaic Homo, with some specimens such as Kabwe, Bodo, and the Ngandong specimens having weaker, and others stonger constriction than Daka. The present study shows that the Daka condition is comparable to (or actually exhibits weaker constriction than) the Jomon/modern H. sapiens condition. Therefore, a marked expression of this measure of constriction is not necessarily primitive for H. sapiens. The results of partial correlation analysis in our Jomon and modern H. sapiens specimens suggest that postorbital constriction relative to the posterior cranium is related not only to endocranial width proportion but also to upper facial breadth. However, as is evident from the strong positive correlation between upper facial breadth and postorbital breadth, a broad upper face associates with reduced expression of constriction. If such effects cut across taxonomic boundaries into archaic Homo in general, so that larger-faced individuals tend to have larger ectocranial measures of postorbital breadth, then this may confound subtle interspecies differences in endocranial width proportions between H. erectus and H. sapiens (Bruner and Holloway, 2010). This may in part account for the weak postorbital constriction relative to maximal cranial breadth in H. erectus (Daka) and early H. sapiens (Skhul V) crania examined in the present study.

Finally, a possible contribution of neuro-orbital disjunction to the expression of constriction was potentially indicated by the significant partial correlation between orbital protrusion and postorbital breadth. Although a significant partial correlation was not obtained between orbital protrusion and postorbital constriction relative to the posterior cranium in our Jomon and modern H. sapiens sample, because of the excessive orbital protrusion values in the Minatogawa crania (especially in Minatogawa I), it is possible that not only endocranial proportions, but also neuro-orbital disjunction in part accounts for their marked postorbital constriction relative to the posterior cranium.

Implications for genealogical relationship

While close morphological and genealogical affinities between the Minatogawa and the Jomon have been suggested and/or assumed (e.g. Suzuki, 1982; Hanihara, 1991; Baba et al., 1998; Brown, 1999), the following distinct cranial features have also been pointed out (Suzuki, 1982; Baba et al., 1998; Kaifu et al., 2011; Saso et al., 2011; Suwa et al., 2011b): the Minatogawa crania have a low position of maximum cranial breadth, a narrow and receding forehead, a swollen glabellosuperciliary prominence associated with a short frontal nasal process, narrow nasal bones, mandibular alveolar prognathism, and large tooth size, while the Jomon people tend to have a high position of the maximum cranial breadth, a broad forehead, bilateral superciliary expression, broad nasals and strong sagittal orientation of the frontal process of the maxilla, tendency for a ‘squarish’ chin, and small tooth crown and root sizes. This morphological contrast suggests a possible genealogical gap between the Minatogawa and the Jomon people. A recent study (Fukase et al., 2012) indicates that the Okinawa skeletons of age equivalent to the Jomon share morphological features such as general facial proportions and broad nasals with the Honshu Jomon.

However, despite Minatogawa’s small endocranial dimensions, when size-standardized by ECV, a certain degree of shared endocranial proportions were found between the Minatogawa and Jomon crania. As it is the case with that of Minatogawa, the Jomon endocranium tends to have a strong temporal bulge compared with the modern Japanese, and PCA summaries place the Jomon mean much closer to Minatogawa than to the modern Japanese mean. In short, in endocranial proportions and morphology, the Minatogawa condition is not so very distinct from the Jomon. On the other hand, the similarity of endocranial proportions between the Minatogawa and Jomon crania does not necessarily support the hypothesis of Minatogawa–Jomon genealogical closeness (Suzuki, 1982; Hanihara, 1991; Baba et al., 1998) because such morphologies were also found in individuals of populations that do not have close genealogical relationships with the Jomon or Minatogawa. However, we here note that our Jomon sample is small considering the extensive time and geographic ranges of the Jomon period, and hence our results should be seen as a starting point for future more extensive comparisons.

Compared to the modern East/Southeast Asians, the modern Australians are thought to retain a significant degree of ancestral craniodental morphologies from the Late Pleistocene populations (Stringer, 1992, 2002). Recent field excavations in Southeast Asia suggest the occurrence of widely distributed Late Pleistocene to mid-Holocene populations with Australo-Melanesian morphological affinities (Matsumura and Hudson, 2005; Matsumura, 2006; Matsumura et al., 2008a, b, 2010). Based on shared mandibular morphologies, Kaifu et al. (2011) and Kaifu and Fujita (2012) suggested that the Minatogawa series might represent a part of a northern expansion of one of such populations in the terminal Pleistocene. On the other hand, in the present study, we observed that a narrow endocranial shape of the modern Australians contrasts with the broad endocranium seen in the Minatogawa crania, especially in Minatogawa I. Lahr and Wright (1996) suggested that modern Australian crania are disproportionately narrow compared with that expected from the relationship between cranial dimensions and robusticity seen among modern populations and some Late Pleistocene specimens, which may indicate that the modern Australian condition is derived. Variation of cranial morphology is undoubtedly in part explained by genetic factors (e.g. Howells, 1973, 1989; Sparks and Jantz, 2002; Relethford, 2004; Manica et al., 2007), but it is also well known that the vault shape of a population can change drastically without large-scale gene flow (e.g. Suzuki, 1969; Kouchi, 2004; Little et al., 2006). We find it essentially difficult to reconstruct genealogical relationships among disparate populations based on external craniofacial and vault or endocranial morphology without a sufficiently dense time successive skeletal record. Discovery of new fossil evidence from the Japanese archipelago, including the Ryukyu Islands as well as from Southeast Asia, is necessary to resolve the significance of the Minatogawa endocranial shape that we have documented in the present study.

Acknowlegdments

The authors thank Dr. N. Ogihara (Keio University) for access to the collections under his care. We also thank Drs. D.E. Lieberman, O. Herschensohn, and M. Morgan (Harvard University) for access to the Skhul V CT image data. We are also grateful to Drs. Y. Kaifu, S. Matsu’ura, M. Kouchi, and O. Kondo for their comments and suggestions. This work was supported in part by Grant-in-Aid for Scientific Research on Innovative Areas (No. 22101001) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

References

•  Adachi  N.,  Shinoda  K.,  Umetsu  K., and  Matsumura  H. (2009) Mitochondrial DNA analysis of Jomon skeletons from the Funadomari site, Hokkaido, and its implication for the origins of Native American. American Journal of Physical Anthropology, 138: 255–265.
•  Aiello  L.C. and  Dean  C. (1990) An Introduction to Human Evolutionary Anatomy. Academic Press, London.
•  Asfaw  B. (1987) The Belohdelie frontal: new evidence of early hominid cranial. Journal of Human Evolution, 16: 611–624.
•  Asfaw  B.,  Gilbert  W.H., and  Richards  G.D. (2008) Homo erectus cranial anatomy. In:  Gilbert  W.H. and  Asfaw  B. (eds.), Homo erectus: Pleistocene Evidence from the Middle Awash, Ethiopia. University of California Press, Berkeley and Los Angeles, pp. 265–327.
•  Baba  H. (2000a) Physical adaptation of the Minatogawa people to island environments. Tropics, 10: 231–241.
•  Baba  H. (2000b) Minatogawa-jin wa Ryukyu-jin no sosen ka: tousho-tekiou no kanten kara [Are the Minatogawa an ancestor of Ryukyuan? From a point of view of the island rule]. In: Collected Papers in Honor of Professor Hiroe Takamiya: Commemoration of His Seventieth Birthday, Ryukyu•higashi Asia no hito to bunka [People and Culture in Ryukyu islands and East Asia], Vol. 2. Syoseido, Nishihara-cho, pp. 413–423.
•  Baba  H. (2002) Minatogawa-jin no ichi-zuke [The positioning of Minatogawa man]. In: Gushichan Village Board of Education (ed.), Minatogawa Fissha Iseki [The Minatogawa Fissure site]. pp. 121–131 (in Japanese).
•  Baba  H. and  Narasaki  S. (1991) Minatogawa Man, the oldest type of modern Homo sapiens in East Asia. The Quaternary Research, 30: 221–230.
•  Baba  H.,  Narasaki  S., and  Ohyama  S. (1998) Minatogawa hominid fossils and the evolution of Late Pleistocene humans in East Asia. Anthropological Science, 106 (Supplement): 27–45.
•  Brown  P. (1999) The first modern East Asians? Another look at Upper Cave 101, Liujiang and Minatogawa 1. In:  Omoto  K. (ed.), Interdisciplinary Perspectives on the Origins of the Japanese. International Research Center for Japanese Studies, Kyoto, pp. 105–131.
•  Bruner  E. and  Holloway  R.L. (2010) A bivariate approach to the widening of the frontal lobes in the genus Homo. Journal of Human Evolution, 58: 138–146.
•  Cameron  D.W. and  Groves  C.P. (2004) Bones, Stones, and Molecules: ‘Out of Africa’ and Human Origins. Elsevier Academic Press, Burlington, MA.
•  Cameron  D.,  Patnaik  R., and  Sahni  A. (2004) The phylogenetic significance of the Middle Pleistocene Narmada hominin cranium from central India. International Journal of Osteoarchaeology, 14: 419–447.
•  Curnoe  D. (2009) Possible causes and significance of cranial robusticity among Pleistocene–Early Holocene Australians. Journal of Archaeological Science, 36: 980–990.
• Editorial Committee of the History of Aichi Prefecture (ed.) (2002) Aichi-ken Shi, Shiryouhen, Vol. 1: Kouko 1 [History of Aichi Prefecture, Materials, Vol. 1: Archaeology 1]. Aichi Prefecture, Nagoya (in Japanese).
•  Fukase  H.,  Mizushima  S.,  Saso  A., and  Suwa  G. (2008) Database of Jomon Period Skeletal Remains 5) Taishaku Kannondo, Taishaku Yosekura. The University Museum, The University of Tokyo, Material Reports No. 73 (in Japanese).
•  Fukase  H.,  Wakebe  T.,  Tsurumoto  T.,  Saiki  K.,  Fujita  M., and  Ishida  H. (2012) Facial characteristics of the prehistoric and early-modern inhabitants of the Okinawa islands in comparison to the contemporary people of Honshu. Anthropological Science, 120: 23–32.
•  Gilbert  W.H. (2008) Hominid systematics. In:  Gilbert  W.H. and  Asfaw  B. (eds.), Homo erectus: Pleistocene Evidence from the Middle Awash, Ethiopia. University of California Press, Berkeley and Los Angeles, pp. 349–371.
•  Gilbert  W.H.,  Holloway  R.L.,  Kubo  D.,  Kono  R.T., and  Suwa  G. (2008) Tomographic analysis of the Daka calvaria. In:  Gilbert  W.H. and  Asfaw  B. (eds.), Homo erectus: Pleistocene Evidence from the Middle Awash, Ethiopia. University of California Press, Berkeley and Los Angeles, pp. 329–347.
•  Hanihara  K. (1991) Dual structure model for the population history of Japanese. Japan Review, 2: 1–33.
•  Hanihara  T. and  Ishida  H. (2009) Regional differences in craniofacial diversity and the population history of Jomon Japan. American Journal of Physical Anthropology, 139: 311–322.
•  Hayashi  K.,  Yamaguchi  B.,  Dodo  Y., and  Hiramoto  Y. (1981) Miyano kaizuka B·C chiku chousa gaiyou—toku ni maisou jinkotsu wo chuushin to shite [Brief summary of the research of region B and C of Miyano shell-mound, with a focus on the buried human skeletal remains]. Report for Grant-in-Aid for Scientific Research from Japanese Ministry on Education (Project No. 331051) (in Japanese).
•  Howells  W.W. (1973) Cranial Variation in Man: A Study by Multivariate Analysis of Patterns of Difference among Recent Human Populations. Papers of the Peabody Museum No. 67. Peabody Museum, Harvard University, Cambridge, MA.
•  Howells  W.W. (1989) Skull Shapes and the Map: Craniometric Analyses in the Dispersion of Modern Homo. Papers of the Peabody Museum No. 79. Peabody Museum, Harvard University, Cambridge, MA.
•  Ikeda  J. (1991) Shuryou-saishuu jidai no hitobito [People of hunting-and-gathering age]. In: Editorial Committee of the History of Okayama Prefecture (ed.), Okayama-ken Shi, Vol. 2: Genshi·Kodai I [History of Okayama Prefecture, Vol. 2: Ancientry]. Okayama Prefecture, Okayama, pp. 103–121 (in Japanese).
•  Prefecture  Iwate (ed.) (1961) Iwate-ken Shi, Vol. 1: Jouko-hen/Joudai-hen [History of Iwate Prefecture, Vol. 1: Ancientry]. Toryo Insatsu, Morioka (in Japanese).
•  Kaifu  Y. and  Fujita  M. (2012) Fossil record of early modern humans in East Asia. Quaternary International, 248: 2–11.
•  Kaifu  Y.,  Aziz  F.,  Indriati  E.,  Jacob  T.,  Kurniawan  I., and  Baba  H. (2008) Cranial morphology of Javanese Homo erectus: new evidence for continuous evolution, specialization, and terminal extinction. Journal of Human Evolution, 55: 551–580.
•  Kaifu  Y.,  Fujita  M.,  Kono  R.T., and  Baba  H. (2011) Late Pleistocene modern human mandibles from the Minatogawa Fissure site, Okinawa, Japan: morphological affinities and implications for modern human dispersals in East Asia. Anthropological Science, 119: 137–157.
•  Kajigayama  M. and  Baba  H. (1999) Human skeletal remains. In:  Bunka Kenkyusho  Rokkou (ed.), Report of Excavation Research of Wakaumi Shell-mound [Wakaumi Kaizuka Hakkutsu Chosa Houkokusho]. Tamatsukuri Town Board of Education, pp. 92–105 (in Japanese).
•  Kimbel  W.H.,  Rak  Y., and  Johanson  D.C. (2004) The Skull of Australopithecus afarensis. Oxford University Press, New York.
•  Kouchi  M. (2004) Secular changes in the Japanese head form viewed from somatometric data. Anthropoligical Science, 112: 41–52.
•  Kubo  D. and  Kono  R.T. (2011) Endocranial restoration and volume estimation of the Minatogawa IV cranium using micro-CT and 3D-printer systems. Anthropological Science, 119: 203–209.
•  Kubo  D.,  Kono  R.T.,  Saso  A.,  Mizushima  S., and  Suwa  G. (2008) Accuracy and precision of CT-based endocranial capacity estimations: a comparison with the conventional millet seed method and application to the Minatogawa 1 skull. Anthropological Science, 116: 77–85.
•  Kubo  D.,  Kono  R.T., and  Suwa  G. (2011) A micro-CT based study of the endocranial morphology of the Minatogawa I cranium. Anthropological Science, 119: 123–135.
•  Kudo  Y. (2004) Reconsidering the geochronological and archaeological framework of the late Pleistocene—early Holocene transition on the Japanese islands. In:  Terberger  T. and  Eriksen  B.V. (eds.), Hunters in a Changing World. Verlag Marie Leidorf, Rahden/Westfalen, pp. 253–268.
•  Lahr  M.M. and  Wright  R.V.S. (1996) The question of robusticity and the relationship between cranial size and shape in Homo sapiens. Journal of Human Evolution, 31: 157–191.
•  Lieberman  D.E. (2000) Ontogeny, homology, and phylogeny in the hominid craniofacial skeleton: the problem of the browridge. In:  O’Higgins  P. and  Cohn  M.J. (eds.), Development, Growth, and Evolution: Implications for the Study of the Hominid Skeleton. Academic Press, London, pp. 85–122.
•  Little  B.B.,  Buschang  P.H.,  Reyes  M.E.P.,  Tan  S.K., and  Malina  R.M. (2006) Craniofacial dimensions in children in rural Oaxaca, southern Mexico: secular change, 1968–2000.
•  Makishima  H. and  Ogihara  N. (2009) Three-dimensional geometric morphometric study of craniofacial variations in Jomon populations. Anthropological Science (Japanese Series), 117: 11–21 (in Japanese with English summary).
•  Manica  A.,  Amos  W.,  Balloux  F., and  Hanihara  T. (2007) The effect of ancient population bottlenecks on human phenotypic variation. Nature, 448: 346–348.
•  Matsumura  H. (2006) The population history of Southeast Asia viewed from morphometric analyses of human skeletal and dental remains. In:  Oxenham  M. and  Tayles  N. (eds.), Bioarchaeology of Southeast Asia. Cambridge University Press, Cambridge, pp. 33–58.
•  Matsumura  H. and  Hudson  M. (2005) Dental perspectives on the population history of Southeast Asia. American Journal of Physical Anthropology, 127: 182–209.
•  Matsumura  H.,  Yoneda  M.,  Dodo  Y.,  Oxenham  M.F.,  Cuong  N.L.,  Thuy  N.K.,  Dung  L.M.,  Long  V.T.,  Yamagata  M.,  Sawada  J.,  Shinoda  K., and  Takigawa  W. (2008a) Terminal Pleistocene human remains skeleton from Hang Cho Cave, northern Vietnam: implications for the biological affinities of Hoabinhian people. Anthropological Science, 116: 201–217.
•  Matsumura  H.,  Oxenham  M.F.,  Dodo  Y.,  Domett  K.,  Thuy  N.K.,  Cuong  N.L.,  Dung  N.K.,  Huffer  D., and  Yamagata  M. (2008b) Morphometric affinity of the late Neolithic human remains from Man Bac, Ninh Binh Province, Vietnam: key skeletons with which to debate the ‘two layer’ hypothesis. Anthropological Science, 116: 135–148.
•  Matsumura  H.,  Domett  K.M., and  O’Reilly  D.J.W. (2010) On the origin of pre-Angkorian peoples: perspectives from cranial and dental affinity of the human remains from Iron Age Phum Snay, Cambodia. Anthropological Science, 119: 67–79.
•  Matsu’ura  S. and  Kondo  M. (2011) Relative chronology of the Minatogawa and the Upper Minatogawa series of human remains from Okinawa Island, Japan. Anthropological Science, 119: 173–182.
•  Mizoguchi  Y. (2011) Typicality probabilities of Late Pleistocene human fossils from East Asia, Southeast Asia, and Australia: implications for the Jomon population in Japan. Anthropological Science, 119: 99–111.
•  Moss  M.L. and  Young  R.W. (1960) A functional approach to craniology. American Journal of Physical Anthropology, 18: 281–292.
•  Nakashima  A.,  Ishida  H.,  Shigematsu  M.,  Goto  M., and  Hanihara  T. (2010) Nonmetric cranial variation of Jomon Japan: implications for the evolution of eastern Asian diversity. American Journal of Human Biology, 22: 782–790.
• R Development Core Team (2008) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, http://www.R-project.org
•  Rak  Y. (1983) The Australopithecine Face. Academic Press, New York.
•  Relethford  J.H. (2004) Boas and beyond: migration and craniometric variation. American Journal of human Biology, 16: 379–386.
•  Rightmire  G.P. (2008) Homo in the Middle Pleistocene: hypodigms, variation, and species recognition. Evolutionary Anthropology, 17: 8–21.
•  Rightmire  G.P. (2009) Middle and later Pleistocene hominins in Afirica and Southwest Asia. Proceedings of the National Academy of Sciences of the United States of America, 106: 16046–16050.
•  Saso  A.,  Matsukawa  S., and  Suwa  G. (2011) Comparative analysis of the glabellar region morphology of the late Pleistocene Minatogawa crania: a three-dimensional approach. Anthropological Science, 119: 113–121.
•  Sparks  C.S. and  Jantz  R.L. (2002) A reassessment of human cranial plasticity: Boas revisited. Proceedings of the National Academy of Sciences of the United States of America, 99: 14636–14639.
•  Stringer  C.B. (1992) Reconstructing recent human evolution. Philosophical Transactions of the Royal Society of London B, 337: 217–224.
•  Stringer  C. (2002) Modern human origins: progress and prospects. Philosophical Transactions of the Royal Society of London B, 357: 563–579.
•  Suwa  G.,  Mizushima  S., and  Sakaue  K. (2003) Database of Jomon Period Skeletal Remains 1) Hobi. The University Museum, The University of Tokyo, Material Reports No. 52 (in Japanese).
•  Suwa  G.,  Fujita  M.,  Yamasaki  S.,  Oshiro  I.,  Baba  H.,  Shinzato  N.,  Kinjo  T.,  Kaifu  Y., and  Matsu’ura  S. (2011a) New insights on the excavation and chronological status of the Late Pleistocene Minatogawa human fossils from Okinawa Prefecture. Anthropological Science (Japanese Series), 119: 125–136 (in Japanese with English summary).
•  Suwa  G.,  Fukase  H.,  Kono  R.T.,  Kubo  D., and  Fujita  M. (2011b) Mandibular tooth root size in the modern Japanese, prehistoric Jomon, and Late Pleistocene Minatogawa human fossils. Anthropological Science, 119: 159–171.
•  Suzuki  H. (1969) Microevolutional changes in the Japanese population from the prehistoric age to the present-day. Journal of the Faculty of Science, University of Tokyo, Section V, 3: 279–308.
•  Suzuki  H. (1982) Skulls of the Minatogawa man. In:  Suzuki  H. and  Hanihara  K. (eds.), The Minatogawa Man: The Upper Pleistocene Man from the Island of Okinawa. The University Museum, The University of Tokyo, Bulletin No. 19, pp. 7–49, plates 2–7.
•  Suzuki  H. and  Hanihara  K. (eds.) (1982) The Minatogawa Man: The Upper Pleistocene Man from the Island of Okinawa. The University Museum, The University of Tokyo, Bulletin No. 19, Tokyo.
•  Takamiya  H. (1996) Initial colonization, and subsistence adaptation processes in the late prehistory of the island of Okinawa. Bulletin of the Indo-Pacific Prehistory Association, 15: 143–150.
•  Weidenreich  F. (1941) The brain and its rôle in the phylogenetic transformation of the human skull. Transactions of the American Philosophical Society, New Series, 31, part 5: 321–442.
•  Wood  B.A. (1991) Koobi Fora Research Project, Vol. 4: Hominid Cranial Remains. Oxford University Press, Oxford.
•  Yamaguchi  B. (1992) Skeletal morphology of the Jomon people. In  Hanihara  K. (ed.), Japanese as a Member of the Asian and Pacific Populations: International Research Symposium No. 4. International Research Center for Japanese Studies, Kyoto, pp. 53–63.
•  Yamasaki  S. and  Takahashi  K. (2007) An Archaeological Study of the Jomon Shellmound at Hikosaki. The University Museum, The University of Tokyo, Bulletin No. 43, Tokyo.
•  Yamasaki  S.,  Fujita  M.,  Ohshiro  I.,  Nishiaki  Y.,  Sugawara  H.,  Eda  M.,  Matsu’ura  S.,  Kondo  M.,  Kaifu  Y.,  Sakaue  K.,  Ozaki  M.,  Mizoguchi  Y.,  Suwa  G., and  Baba  H. (2010) Okinawa ni okeru Koshinsei/Kanshinsei ikoki no jinruigaku-kokogaku chosa [Anthropological and archaeological research on the Pleistocene/Holocene transitional period in Okinawa]. Abstracts with Programs, The 76th Annual Meeting of the Japanese Archaeological Association, pp. 120–121 (in Japanese).
•  Yoneda  M.,  Saso  A.,  Suzuki  R.,  Shibata  Y.,  Morita  M.,  Suwa  G., and  Akazawa  T. (2005) Chronology of the Yayoi skeletal remains from the Kanto district, Japan: a preliminary reevaluation by radiocarbon dating of postcranial material. Anthropological Science, 113: 169–182.