Anthropological Science Vol. 128(2), 71–81, 2020

Subspecies and sexual craniofacial size and shape variations in Japanese macaques (Macaca fuscata) Wataru Yano1*, Naoko Egi2, Tomo Takano3, Naomichi Ogihara4 1Laboratory of Biology, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan 2Phylogeny and Systematics section, Primate Research Institute, Kyoto University, Inuyama, Aichi 484-8506, Japan 3Japan Monkey Centre, 26 Kanrin, Inuyama, Aichi 484-0081, Japan 4Laboratory of Human Evolutionary Biomechanics, Department of Biological Sciences, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan

Received 16 December 2018; accepted 5 May 2020

Abstract Adaptation to various environments leads to evolutionary change in size and shape in non- human primates. In island environments, larger mammals tend to be smaller compared with the original mainland population. The Japanese macaque (Macaca fuscata) has two subspecies: Macaca fuscata yakui (MFY) on Yakushima Island, and Macaca fuscata fuscata (MFF) on the main Japanese archipelago. Since adult shape differences reflect spatiotemporal developmental pattern differences, it is interesting to examine allometric patterns for groups that show significant size variation. The main purpose of the present study is to quantitatively examine the craniofacial size and three-dimensional sexual and subspecies shape variation, focusing on the effects of size variation on shape variation. Computed to- mography scans of 55 specimens were used to generate a three-dimensional model of each cranium, and 57 landmarks were digitized to quantify the craniofacial shape variation in Japanese macaques. We subsequently employed regression analyses to deduce vectors responsible for allometry, sex, subspecies shape variations, and the influence of size on sexual and subspecies shape variations. The results showed that four intraspecific groups, consisting of two subspecies and the two sexes, significantly differed in both size and shape space. In size, the cranium of MFY was smaller than that of MFF in both sexes, and female crania were smaller than male crania in both subspecies. Allometry as well as sexual dimorphism in shape was related to a relatively broad orbit, smaller neurocranium, enlarged snout, and broader temporal fossa in males. Subspecies shape differences were a relatively narrow and short orbit and sphenoid, smaller neurocranium, and postorbital constriction in MFY. Sexual shape variation was largely associated with size variation. On the other hand, subspecies shape variation was not significantly correlated with size. We discuss these intraspecific cranial size and shape variations and the effect of size on shape variation from evolutional and developmental perspectives.

Key words: Macaca fuscata, geometric morphometrics, subspecies, sexual dimorphism, allometry

(MFF). Later, molecular phylogenetic analyses demonstrated Introduction that the genetic variation of MFY falls within the range of The Japanese macaque (Macaca fuscata) has the most the MFF population and the subspecific status of MFY has northerly range among extant non-human primates, and is consequently been questioned (blood protein: Nozawa et al., endemic to the Japanese archipelago. Kuroda (1940) first 1991; mtDNA: Marmi et al., 2004; Kawamoto et al., 2007; distinguished a southern population on Yakushima Island Kawamoto, 2010). On the other hand, these molecular (504.9 km2) from other populations based on its reduced analyses revealed the isolation and subsequent unique body size and diagnostic pelage color, and he proposed sub- evolutionary history of MFY. According to Nozawa et al. species recognition of Macaca fuscata yakui (MFY), distinct (1991), MFY and MFF diverged 178 kya ago, and gene flow from their mainland counterpart Macaca fuscata fuscata from the mainland of the Japanese archipelago to Yakushima Island after Marine Isotope Stages (MIS) 6–8 (130–300 kya) is unlikely. Hayaishi and Kawamoto (2006) also showed that * Correspondence to: Wataru Yano, Laboratory of Biology, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359- there should have been no migration between Yakushima 8513, Japan. and the nearest Kyushu Island of the Japanese mainland. E-mail: [email protected] After divergence, about 7300 years ago the population in Published online 31 July 2020 Yakushima Island suffered from a bottleneck effect due to a in J-STAGE (www.jstage.jst.go.jp) DOI: 10.1537/ase.2005052 massive volcanic explosion 40 km from this island (Hayaishi

© 2020 The Anthropological Society of Nippon 71 72 W. YANO ET AL. Anthropological Science and Kawamoto, 2006). of isolated primates in island environments. We discuss the Cranial form differences between MFY and MFF have evolutional and developmental factors responsible for the been diagnosed by using linear distance-based morphometrics variation in cranial size and shape in the Japanese macaque. (craniometry: Ikeda and Watanabe, 1966; Kuroda 1984, 2002; Mouri and Nishimura, 2002; somatometry: Iwamoto, Methods 1971; Hamada et al., 1996). According to these studies on size variation, MFY crania are generally smaller than MFF We carefully inspected cranial specimens to exclude crania. Regarding shape, Ikeda and Watanabe (1966) pointed broken and/or pathological samples, and a total of 55 adult out that MFY has distinct cranial features with expanded dried crania of Japanese macaque (14 male MFFs, 13 female zygoma, postorbital constriction, higher lambda and inion, MFFs, 14 male MFYs, and 14 female MFYs) were obtained narrower orbit, and a protrusive snout. As for sexual shape from the Laboratory of Physical Anthropology, Kyoto Uni- variation, Ikeda and Watanabe (1966) reported that the versity (LPA; Kyoto, Japan), the Primate Research Institute, crania of male Japanese macaques showed a relatively large Kyoto University (PRI; Inuyama, Japan), and the Japan splanchnocranium. However, these studies did not conduct Monkey Centre (JMC; Inuyama, Japan). Since wild and further statistical analysis on the relationship between size captive Japanese macaques do not differ significantly and shape. This is partly due to limitations in conventional (Kamaluddin et al., 2019), we employed mixed samples of morphometrics for quantifying biological form. Linear wild and captive origin as well as a few samples of unknown measurements involve removing any spatial information origin from animal traders. We carefully used unbiased other than the linear distance between two landmarks. Since numbers for the different groups (Table 1). For wild the positional data of each landmark are omitted, the observ- specimens of MFF, we sampled dried crania from a wide er cannot maintain relative spatial relationships (i.e. variety of populations inhabiting the mainland Japanese ar- geometry) between landmarks. For example, even though chipelago. We used only adult specimens who have upper we can detect the increase in cranial length by linear meas- third molars fully erupted. urement, we cannot know if the glabella and/or inion chang- Each specimen was scanned with a computed tomography es their positions. Linear measurement also has limitations (CT) scanner (TSX-002A/4I; Toshiba Medical Systems, in extensive statistical analyses. Once decomposed into a set Tokyo, Japan) in LPA. The tube voltage and current were set of linear dimensions, the integrity of the biological form at 120 kV and 100 mA. Cross-sectional images were recon- decays and subsequent statistical analyses with external structed with a pixel size of 0.20–0.25 mm and a slice inter- variables or visualization of the statistical results can be val of 0.20 mm. The 3D surface of the cranium was then partial and incomplete (Zelditch et al., 2004; Slice, 2005). generated using a triangular mesh model with commercial To rigorously quantify subspecies shape variations, and software (Analyze 6.0; Mayo Clinic, Rochester, MN, USA). the relationship of size and shape variation, we employed We digitized a total of 57 landmarks (Figure 1, Table 2) landmark-based three-dimensional geometric morphomet- on the external and internal surfaces of each cranium using rics (GM). GM enables integrated quantification of biologi- commercial software (Rapidform 2004; INUS Technology, cal form without losing geometric information of landmarks. Seoul, Korea). All the crania were digitized by the same We employed generalized Procrustes analysis (GPA) person (W.Y.). As shape variances due to left–right asymme- (Zelditch et al., 2004; Slice, 2005). In this morphometry, try were not considered here, we symmetrized the positions landmark sets of samples are first normalized, translated to of all landmarks using self-symmetrization (Zollikofer and the same centroid position, scaled into the same centroid Ponce de León, 2002). Specifically, we created the horizon- size, and finally rotated about the centroid using least- tally reflected specimen for each sample and superimposed squares fitting. The coordinate data of landmark sets after them by least-squares superposition to calculate the mean of fitting denotes its intrinsic shape. Since the original centroid each landmark coordinate, yielding the symmetrized speci- position, centroid size, and rotation angle about centroid are men of the original. retrievable, we can use them as external parameters as well. Then we analyzed the variances in the landmark positions Furthermore, since the composition of a set of landmarks is using the geometric morphometric software Morphologika unchanged, we can visualize any statistical procedure as the version 2.3.1 (O’Higgins and Jones, 1998: http://sites.goog- deformation of landmarks (Zelditch et al., 2004; Slice, le.com/site/hymsfme/resources). In this method, centroid 2005). The usefulness of this method has been tested in size (CS) was calculated from the coordinates of each crani- inter- and intraspecific comparison of size and shape in um as the square root of the sum of squared Euclidean dis- primate crania (e.g. O’Higgins and Dryden, 1993; Frost et tances from each landmark to the mean of the configuration al., 2003; Pan et al., 2003; Schaefer et al., 2004; Cardini and of landmark coordinates. CS can be used as a proxy size Elton, 2008a, b, 2017; Ito et al., 2014). Here, we first aim to clarify through the use of GM the intraspecies cranial size and shape variations among four Table 1. Numbers of wild/captive/unkown specimens groups of Japanese macaque consisting of two subspecies Wild Captive Unknown and the two sexes. Second, we aim to clarify the effect of size on sexual dimorphism and subspecific difference, MFF male 12 1 1 respectively. The influence of cranial size reduction on MFF female 7 2 4 subspecific differences deserves special consideration MFY male 10 3 1 because it can shed light on the peculiar dwarfing evolution MFY female 10 3 1 Vol. 128, 2020 CRANIOFACIAL SIZE AND SHAPE VARIATIONS IN MACACA FUSCATA 73

Figure 1. Landmarks and wireframe used in the present study. (A) Anterior view. (B) Lateral view. (C) Posterior view of sphenoid. (D) Inferior view.

variable for each sample. Subsequently, the landmark con- permutation testing (n = 1000). The shape variations were figurations of each cranium were standardized with GPA to visualized using 3D deformation of the wireframe connect- be projected onto a single point in Kendall’s shape space ing the landmarks. These analyses were done with MorphoJ (Zelditch et al., 2004; Slice, 2005). Then, the points were 1.07a (Klingenberg, 2011). projected onto a linear tangent space for subsequent statisti- cal analyses. Principal component analysis (PCA) was con- Results ducted to extract principal components (PCs) of shape vari- ations among crania (O’Higgins and Jones, 1998). Within The four intraspecific groups clearly differed from one this shape space, variations of the four intraspecific groups another in size variation represented by CS (Figure 2A). The were compared. We first explored the PC1–15 axes, which ANOVA showed that mean scores of CS for the four in- showed inter-group differences. We tested deviation from traspecific groups were not homogeneous (P < 0.001), and the normal distribution of CS and PC scores for each group subsequent post-hoc Tukey’s HSD test revealed that males with the Shapiro–Wilk test. When this normality test was had larger CS than females for both subspecies (P < 0.001), passed, analysis of variance (ANOVA) tests and post-hoc and likewise, MFF were larger than MFY for both sexes Tukey’s HSD tests were employed to test for significant dif- (P < 0.01) (Figure 3A). The shape variations are illustrated ferences in CS and the PC scores among the four intraspecif- as scatter plots of PC1 versus PC2, and PC1 versus PC3, in ic groups. These statistical tests were performed with Statis- Figure 3. Along the PC1 axis, males clearly had higher tica 2000 version 5.5 (StatSoft, Tulsa, OK, USA). scores than females for both subspecies. Likewise, MFY had Subsequently, we explored axes denoting allometry, slightly higher PC1 scores than MFF for both sexes. The sexual dimorphism, and subspecific differences, by multiple ANOVA test demonstrated that mean PC1–3 scores for the regression analyses with CS, as well as dummy codes (0/1) four groups were not the same (P < 0.001). Tukey’s HSD for sex and subspecies, respectively. Next, these regression test revealed pairwise differences between the four groups. scores for each sexual and subspecific shape variations were In PC1, males had higher scores than females for both sub- further subject to correlation and regression analyses by size species (P < 0.001), and MFY had higher scores than MFF and PC axes, in order to explore their contribution. Each for both sexes (P < 0.05) (Figure 2B, Figure 3A). In PC2, regression analysis was conducted with the ordinary least- scores decreased in the order of MFF male, MFF female, squares method and the significance of the coefficient of MFY male, and MFY female, whereas significant differences slope in each regression analysis was calculated from were found between MFF males and MFY males, and MFF 74 W. YANO ET AL. Anthropological Science

Table 2. Landmarks used in this study No. Definition Type 1 Bregma M 2 Supraorbitale M 3 Maxillofrontale B 4 Most superior point on the inferior orbital rim of frontal bone B 5 Frontomalare orbitale B 6 Sphenion B 7 Ectocranial midline point on superior rim of anterior nasal aperture M 8 Alare B 9 Nasospinale M 10 Frontomalare temporale B 11 Zygoorbitale B 12 Most inferolateral point on inferior orbital rim B 13 Prosthion M 14 Jugale B 15 Zygomaxillare B 16 Most posterior point in the temporal process of the zygomatic bone B 17 Point in the depth of the angle between the zygomatic process and the squama of temporal bone B 18 Midpoint on the buccal alveolar rim of second molar B 19 Staphylion M 20 Opisthion M 21 Inion M 22 Ectocranial midline point where inferior nuchal line crosses M 23 Asterion B 24 Auriculare B 25 Cross-sectional point of median and transverse suture of palatine M 26 Sphenobasion M 27 Most anterior point of the hypoglossal canal of lateral part of occipital bone B 28 Basion M 29 Most lateral point on the lateral margin of the foramen magnum of lateral part of occipital bone B 30 Most lateral point on the optic canal of lesser wing of sphenoid bone B 31 Midpoint on the lateral side of the superior surface of the postsphenoid part of the body of sphenoid bone B 32 Most superolateral point on the grater wing of sphenoid bone B 33 Most inferolateral point on the greater wing of sphenoid bone B 34 Most inferior point on the foramen rotundum of sphenoid bone B 35 Nasion M M, midsagittal; B, bilateral. females and MFY females (P < 0.001) (Figure 2C, (Figure 4A). Regarding regression for shape variations Figure 3A). In PC3, only MFF males and MFY females using dummy codes (0/1), we successfully deduced sexual differed significantlyP ( < 0.01) (Figure 2D, Figure 3B). shape variation (F(1,82) = 5.264, P = 0.0002) and subspecies The proportions of eigenvalues of PCA are listed in shape variation (F(1,82) = 14.07, P < 0.0001). For the effect Table 3. Approximately 80% of the variance was incorporat- of size on shape variation, regression of size on sexual ed in the first 15 PCs. Significant differences among the four dimorphism (SD) was significant (r = 0.645, P < 0.0001) intraspecific groups were found for PC1, PC2, and PC3 with (Figure 4B, Table 5, Table 6). On the other hand, regression the ANOVA test. In other axes, the positions of four groups of size on subspecies shape variation (SS) was not significant overlapped each other. Among PC1–15, PC1, PC2, and PC5 (r = –0.134, P = 0.329) (Figure 4C, Table 5, Table 6). were significantly correlated with CS (Table 4). PC1 was We constructed visual presentations of 3D shape variations also strongly correlated with sexual dimorphism (Figure 3A, among the four intraspecific groups due to changes along the Table 5), and about 95% of shape variation between sex can allometric (Figure 5), sexual (Figure 6), and subspecific be explained on PC1 (Table 6). Subspecies shape variation shape (Figure 7) axes. Cranial shape was represented by the correlated with PC1, PC2, and PC3, while the contributions wireframe connecting landmarks. Three-dimensional shape of these three axes were heterogeneous (PC1, 38.82%; PC2, variation along the allometric axis is shown in Figure 5. 37.90%; PC3, 17.50%). Crania of larger samples, having positive regression scores, The results of regression analyses are shown in Figure 4. were characterized by a lower neurocranium, a First, cranial shape variation of Japanese macaque was superoanteriorly positioned as well as vertically tilted significantly correlated with size (r = 0.811, P < 0.0001) nuchal crest, a mediolaterally expanded zygomatic arch, a Vol. 128, 2020 CRANIOFACIAL SIZE AND SHAPE VARIATIONS IN MACACA FUSCATA 75

Figure 2. Comparison of mean scores among the four intraspecific groups. (A) Centroid size (B) PC1 (C) PC2 (D) PC3. Whisker plots represent 95% confidence intervals. P* < 0.05, **P < 0.01, ***P < 0.001. shorter orbit height, and a relatively developed face with to sexual dimorphism of size, and differences in the protruded muzzle compared to the neurocranium. Sexual adolescent spurt pattern result in differences in adult cranial shape variation, which was highly correlated with size, was size, especially in the facial region. In the Japanese ma- similar to the allometric variation, having a smaller caque, although the spurt starts earlier in females, growth neurocranium and developed face with protruded muzzle as rate and duration are larger in males, which results in a high well as a larger zygomatic arch and superoinferiorly shorter total amount of growth and the male bearing a larger orbits (Figure 6). On the other hand, in subspecies shape splanchocranium (Hamada, 1994; Mouri, 1994). Regarding variation, MFY tended to have a lower neurocranium, subspecies size differences, the difference in developmental relatively vertically tilted nuchal crest, stronger postorbital processes between subspecies is not fully understood. constriction, and narrower sphenoid and orbit (Figure 7). According to Hamada (1994), MFY is smaller at birth. Since our previous study demonstrated cranial shape variation in Discussion MFY with a relatively small neurocranium, suggesting the truncation or retardation of fetal neurocranium growth in This study clarified distinctive cranial size and 3D shape MFY (Yano et al., 2010), it is possible that fetal neurocranial variations among four groups consisting of two subspecies growth is important in the development of the size difference and the two sexes of Macaca fuscata. The significant size between the subspecies MFY and MFF. differences in sex (female < male) and subspecies Distinct shape differences were also found between sub- (MFY < MFF) are consistent with previous studies (Ikeda species and between sexes (Figure 3). Since sexual and Watanabe, 1966; Iwamoto, 1971; Hamada et al., 1996; dimorphism was almost similar between the two subspecies, Mouri and Nishimura, 2002). The pubertal period is pivotal the common sexual dimorphism was illustrated as a shape 76 W. YANO ET AL. Anthropological Science

Figure 3. Sexual dimorphism and subspecies variation in cranial shape of Macaca fuscata. Plots of principal component analysis: (A) PC1 vs. PC2. (B) PC1 vs. PC3. Sexual dimorphism for each subspecies and subspecies variation for each sex are indicated by dashed and solid lines, respectively.

deformation of the wireframe (Figure 6). Compared with depending on their different intrinsic developmental patterns females, males had a smaller neurocranium, a developed (Mitteroecker et al., 2004; Bastir and Rosas, 2009). face with a protruded muzzle, larger zygomatic arch, and a Modification of growth patterns among modules can be superoinferiorly shorter orbit; these are similar to the responsible for shape variation among adults. For sexual findings in other primates (e.g. Cardini and Elton, 2008b, shape dimorphism, during the adolescent period, males ex- 2017). Shape variation in adults can be understood as the perience a large amount of growth in the facial region result of different growth proportions in each component of (Mouri, 1994). Relative enlargement of the lower face in biological form. We can deconstruct the cranium into males is associated with canine development, which is req- modules such as the face, basicranium, and neurocranium uisite for male–male competition (Plavcan, 2001). Other Vol. 128, 2020 CRANIOFACIAL SIZE AND SHAPE VARIATIONS IN MACACA FUSCATA 77

Table 3. Proportion of eigenvalues of principal components accommodating the insertion of nuchal muscles, and a Axis Variance (%) Cumulative (%) broader temporal fossa for temporal muscle passage. These PC1 20.0 20.0 sexual shape differences were consistent with the findings of PC2 11.7 31.7 a previous study using linear morphometrics (Ikeda and PC3 9.36 41.1 Watanabe, 1966). Since sexual shape variation is correlated PC4 6.17 47.3 with size, shape variation can be largely explained as a PC5 5.54 52.8 consequence of the size differences between males and PC6 5.04 57.8 females. To test heterochronic processes such as ontogenetic PC7 3.78 61.6 scaling between sexes, additional data from infants and PC8 3.42 65.0 adolescents is required to elucidate postnatal ontogenetic PC9 2.87 67.9 trajectories (e.g. Cobb and O’Higgins, 2007). PC10 2.79 70.7 Among subspecies shape differences, the vertically tilted PC11 2.43 73.1 nuchal crest, postorbital constriction, and narrower orbit demonstrated in MFY are concordant with the differences PC12 2.05 75.2 reported by Ikeda and Watanabe (1966). We also found additional subspecies differences, such as a narrower Table 4. Correlation coefficients between CS and PCs sphenoid in MFY. In terms of the relationship between size and shape, size does not contribute to the cranial shape r P variation between subspecies (Figure 4C). Although it was PC1 0.533 <0.0001 **** not significant, the regression slope indicates that many parts PC2 0.463 0.0002 *** of the characteristic traits of MFY, such as a smaller PC3 0.173 0.108 neurocranium, postorbital constriction, and more vertically PC4 0.136 0.891 tilted nuchal plane, are generally the characteristics found in PC5 0.274 0.024 * larger individuals in macaques (Ito et al., 2014). Kuroda PC6 0.133 0.819 (1984, 2002) also pointed out this phenomenon in his non- *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. metric study, noting that MFY is more hyperostetic than MFF. Since it is not likely that smaller-sized MFY grow more than MFF, it is feasible that evolution for a smaller sexually dimorphic characters are associated with the rela- neurocranium during fetal development makes MFY have a tive development of muscles attached to the external surface proportionally larger face. Since the neurocranium develops of the cranium, such as a higher and broader nuchal plane mostly in utero, modification of the developmental pattern in

Table 5. Correlation coefficients between CS and PCs and sexual dimorphism (SD) and subspecific shape diffrence (SS) Sexual dimorphism (SD) Subspecific difference (SS) r P r P CS 0.645 <0.0001 *** CS –0.134 0.330 PC1 0.975 <0.0001 *** PC1 0.645 <0.0001 **** PC2 0.210 0.1251 PC2 –0.599 <0.0001 **** PC3 0.021 0.8855 PC3 –0.411 0.0017 ** PC4 –0.003 0.9806 PC4 –0.004 0.9787 PC5 –0.017 0.9035 PC5 –0.173 0.2068 PC6 0.037 0.7927 PC6 –0.008 0.9556 *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Table 6. Regression coefficients of regression analyses by CS and PCs on sexual dimorphism (SD) and subspecific shape diffrence (SS) Sexual dimorphism (SD) Subspecific difference (SS) Slope P Predicted Slope P Predicted CS 0.0008 <0.0001 **** 42.19% CS –0.0002 0.2492 2.48% RS **** 96.01% RS 1.071 <0.0001 **** 96.01% PC1 0.944 <0.0001 **** 94.72% PC1 0.478 <0.0001 **** 38.82% PC2 –0.275 0.1104 4.73% PC2 0.616 <0.0001 **** 37.90% PC3 –0.036 0.8598 0.07% PC3 0.469 0.015 ** 17.50% PC4 0.006 0.9803 0.00% PC4 0.005 0.9789 0.00% PC5 –0.028 0.9117 0.02% PC5 –0.255 0.1973 3.06% PC6 –0.072 0.7802 0.14% PC6 0.014 0.9459 0.01% *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. 78 W. YANO ET AL. Anthropological Science

Figure 4. Contribution of size to shape variation. Regression analyses by centroid size (CS) for (A) general shape variation (allometry), (B) sexual dimorphism in shape, (C) subspecific difference in shape (SS). this period would be essential for the cranial shape evolution and Zollikofer, 2001; Ackermann and Krovitz, 2002; Cobb of MFY from MFF. Yano et al. (2010) indicated that and O’Higgins, 2004). divergence of ontogenetic trajectories occurs at a very early From an ecological perspective, a smaller neurocranium stage of fetal life. The early divergence of ontogenetic may be adaptive for saving energy in an island environment trajectories has also been suggested in human and non- (Foster, 1964; Lomolino, 1985). It is also notable that the human primates (Richtsmeier et al., 1993; Ponce de León environment of MFY has been vulnerable to natural Vol. 128, 2020 CRANIOFACIAL SIZE AND SHAPE VARIATIONS IN MACACA FUSCATA 79

Figure 5. Shape variation in terms of allometry. Shape variations were visualized with 3D deformation of the wireframe connecting landmarks. Solid lines, large; dashed lines, small. Scaling factor was set to 50, representing the difference between male and female in CS.

Figure 6. Visualization of sexual dimorphism in shape. The vector for shape change was explored by regression of shape variation with a dummy variable (male = 1, female = 0). Solid lines, males; dashed lines, females. Scaling factor was set to 1, denoting Procrustes distance either side of the mean shapes. disasters. Specifically, the MFY population was inbreeding and homozygosity. According to a recent human catastrophically damaged after devastating volcanic genomic study (Joshi et al., 2015), increased homozygosity pyroclastic flows about 7300 years ago (Hayaishi and can reduce stature height. Previous somatometric studies Kawamoto, 2006). Low genetic diversity can increase demonstrated that MFY have significantly shorter head and 80 W. YANO ET AL. Anthropological Science

Figure 7. Visualization of subspecies differences in shape. Regression analysis of shape variation by dummy variable. The vector for shape change was explored by regression of shape variation with a dummy variable (MFY = 1, MFF = 0). Solid lines, MFY; dashed lines, MFF. Scaling factor was set to 1, denoting dummy code difference. body lengths than MFF (Iwamoto, 1971; Hamada et al., embryogenesis of the cranium is responsible for the 1996; Buck et al., 2018). Different from other cranial bones, morphological evolution in MFF and MFY. basicranial bones, such as the occipital base and sphenoid body, develop in endochondral ossification like the bones of Acknowledgments the vertebral column. Influenced in a similar way by hormones such as the pituitary growth hormone, an We wish to sincerely thank Kazumichi Katayama, Masato association between somatic and cranial base development Nakatsukasa, Toshisada Nishida, and Daisuke Shimizu for has been reported (Cantu et al., 1997; Nie, 2005). their continuous guidance and support throughout the course Accordingly, evolutionary change for a shorter body of the present study. This study was supported by Japan (dwarfism) can heterogeneously affect the cranium, Society for the Promotion of Science (JSPS) Grant-in-Aid specifically showing greater effects on the basicranial bones. for Scientific Research (B) 19370101 to N.O. and in part by In fact, in the present study, reduction in the size of the orbit the Global Center of Excellence Program A06 ‘Formation of and the sphenoid bone in MFY were indicated by a relatively a Strategic Base for Biodiversity and Evolutionary Re- narrow and shorter orbital and sphenoid region. Complete search: from Genome to Ecosystem’ of the Ministry of Edu- analysis the evolutionary mechanisms involved in the cation, Culture, Sports, Science and Technology (MEXT), cranial shape of these island primates will require further Japan. We are very grateful to the two anonymous reviewers morphometric studies to quantify different development for their comments and suggestions to improve this patterns of cranial bones and to integrate these into the manuscript. overall cranial morphology. References Conclusion Ackermann R.R. and Krovitz G.E. (2002) Common patterns of fa- cial ontogeny in the hominid lineage. Anatomical Record, The present study demonstrated significant 3D 269: 142–147. morphological variation in crania of Japanese macaque. The Bastir M. and Rosas A. (2009) Mosaic evolution of the basicranium intraspecies group consists of two sexes and two subspecies, in Homo and its relation to modular development. Evolutionary which clearly differed both in size and shape. The present Biology, 36: 57–70. data showed that sexual shape dimorphism can be explained Buck L.T., De Groote I., Hamada Y., and Stock J.T. (2018) Humans preserve non-human primate pattern of climatic adaptation. by size difference, indicating that source of shape variation Quaternary Science Reviews, 192: 149–166. originates from the elongation of growth in males. On the Cantu G., Buschang, P.H., and Gonzalez J.L. (1997) Differential other hand, subspecies shape variation was not associated growth and maturation in idiopathic growth-hormone- with size variation, suggesting that genetic and/or early deficient children. European Journal of Orthodontics, 19: Vol. 128, 2020 CRANIOFACIAL SIZE AND SHAPE VARIATIONS IN MACACA FUSCATA 81

131–139. Yakushima macaques. Monkey, 198: 14–17 (in Japanese). Cardini A. and Elton S. (2008a) Variation in guenon skulls (I): spe- Kuroda S. (2002) Microevolution of cranial bone morphology of cies divergence, ecological and genetic differences. Journal of Macaca fuscata in the postglacial period non-metrological Human Evolution, 54: 615–637. mutant characters of cranial bone of Macaca fuscata. Asian Cardini A. and Elton S. (2008b) Variation in guenon skulls (II): Paleoprimatology, 2: 115–125 (in Japanese). sexual dimorphism. Journal of Human Evolution, 54: 638– Lomolino M.V. (1985) Body size of mammals on islands: the is- 647. land rule reexamined. American Naturalist, 125: 310–316. Cardini A. and Elton S. (2017) Is there a ‘Wainer’s rule’? Testing Marmi J., Bertranpetit J., Terradas J., Takenaka O., and Domingo- which sex varies most as an example analysis using GueSDat, Roura X. (2004) Radiation and phylogeography in the the free Guenon Skull Database. Hystrix, 28147–28156. Japanese macaque, Macaca fuscata. Molecular Phylogenetics Cobb S.N. and O’Higgins P. (2004) Hominins do not share a com- and Evolution, 30: 676–685. mon postnatal facial ontogenetic shape trajectory. Journal of Mitteroecker P., Gunz P., Bernhard M., Schaefer K., and Bookstein Experimental Zoology Part B: Molecular and Developmental F.L. (2004) Comparison of cranial ontogenetic trajectories Evolution, 302: 302–321. among great apes and humans. Journal of Human Evolution, Cobb S.N. and O’Higgins P. (2007) The ontogeny of sexual 46: 679–698. dimorphism in the facial skeleton of the African apes. Journal Mouri T. (1994) Postnatal growth and sexual dimorphism in the of Human Evolution, 53: 176–190. skull of the Japanese macaque (Macaca fuscata). Anthropo- Foster J.B. (1964) Evolution of mammals on islands. Nature, 202: logical Science, 102 Supplement: 43–56. 234–235. Mouri T. and Nishimura T. (2002) Craniometry of adult male japa- Frost S.R., Marcus L.F., Bookstein F.L., Reddy D.P., and Delson E. nese macaques from the Yakushima, Koshima and Kinkazan (2003) Cranial allometry, phylogeography, and systematics of islands. Primate Research, 18: 43–47 (in Japanese with Eng- large-bodied papionins (primates: Cercopithecinae) inferred lish summary). from geometric morphometric analysis of landmark data. An- Nie X. (2005) Cranial base in craniofacial development: atomical Record Part A: Discoveries in Molecular, Cellular, developmental features, influence on facial growth, anomaly, and Evolutionary Biology, 275: 1048–1072. and molecular basis. Acta Odontologica Scandinavica, 63: Hamada Y. (1994) Standard growth patterns and variations in 127–135. growth patterns of the Japanese monkeys (Macaca fuscata) Nozawa K., Shotake T., Minezawa M., Kawamoto Y., Hayasaka K., based on an analysis by the spline function method. and Kawamoto S. (1991) Population genetics of Japanese Anthropological Science, 102 Supplement: 57–76. monkeys: III Ancestry and differentiation of local populations. Hamada Y., Watanabe T., and Iwamoto M. (1996) Morphological Primates, 32: 411–435. variations among local populations of Japanese macaque O’Higgins P. and Dryden I.L. (1993). Sexual dimorphism in homi- (Macaca fuscata). In: Shotake T. and Wada K. (eds.), Varia- noids: further studies of craniofacial shape differences in pan, tions in the Asian Macaques. Tokai University Press, Tokyo, gorilla and pongo. Journal of Human Evolution, 24: 183–205. pp. 97–115. O’Higgins P. and Jones N. (1998) Facial growth in Cercocebus Hayaishi S. and Kawamoto Y. (2006) Low genetic diversity and torquatus: an application of three-dimensional geometric mor- biased distribution of mitochondrial DNA haplotypes in the phometric techniques to the study of morphological variation. Japanese macaque (Macaca fuscata yakui) on Yakushima Is- Journal of Anatomy, 193: 251–272. land. Primates, 47: 158–164. Pan R., Wei F., and Li M. (2003) Craniofacial variation of the Chi- Ikeda J. and Watanabe T. (1966) Morphological studies of Macaca nese macaques explored with Morphologika. Journal of Mor- fuscata. Primates, 7: 271–288. phology, 256: 342–348. Ito T., Nishimura T., and Takai M. (2014) Ecogeographical and Plavcan J.M. (2001) Sexual dimorphism in primate evolution. phylogenetic effects on craniofacial variation in macaques. Yearbook of Physical Anthropology, 44: 25–53. American Journal of Physical Anthropology, 154: 27–41. Ponce de León M.S. and Zollikofer C.P.E. (2001) Neanderthal cra- Iwamoto M. (1971) Morphological studies of Macaca fuscata: VI. nial ontogeny and its implications for late hominid diversity. Somatometry. Primates, 12: 151–174. Nature, 412: 534–538. Joshi P.K., Esko T., Mattsson H., Eklund N., Gandin I., et al. (2015) Richtsmeier J.T., Corner B.D., Grausz H.M., Cheverud J.M., and Directional dominance on stature and cognition in diverse Danahey S.E. (1993) The role of postnatal growth pattern in human populations. Nature, 523: 459–462. the production of facial morphology. Systematic Biology, 42: Kawamoto Y. (2010) Modes of differentiation in Japanese 307–330. macaques: perspectives from population genetics. In: Schaefer K., Mitteroecker P., Gunz P., Bernhard M., and Bookstein Nakagawa N., Nakamichi M., Sugiura H. (eds.), The Japanese F.L. (2004) Craniofacial sexual dimorphism patterns and al- Macaques. Springer, Tokyo, pp. 53–76. lometry among extant hominids. Annals of Anatomy/Anato- Kawamoto Y., Shotake T., Nozawa K., Kawamoto S., Tomari K.I., mischer Anzeiger, 186: 471–478. et al. (2007) Postglacial population expansion of Japanese Slice D. (2005) Modern Morphometrics in Physical Anthropology. macaques (Macaca fuscata) inferred from mitochondrial Kluwer Academic and Plenum, New York. DNA phylogeography. Primates, 48: 27–40. Yano W., Egi N., Takano T., and Ogihara N. (2010) Prenatal ontog- Kamaluddin S.N., Tanaka M., Wakamori H., Nishimura T., and Ito eny of subspecific variation in the craniofacial morphology of T. (2019) Phenotypic plasticity in the mandibular morphology the Japanese macaque (Macaca fuscata). Primates, 51: 263– of Japanese macaques: captive–wild comparison. Royal Soci- 271. ety Open Science, 6: 181382. Zelditch M., Swiderski D., Sheets H., and Fink W. (2004) Geomet- Klingenberg C.P. (2011) MorphoJ: an integrated software package ric Morphometrics for Biologists: A Primer. Elsevier, San for geometric morphometrics. Molecular Ecology Resources, Diego. 11: 353–357. Zollikofer C.P.E. and Ponce de León M.S. (2002). Visualizing pat- Kuroda N. (1940) A Monograph of the Japanese Mammals exclu- terns of craniofacial shape variation in Homo sapiens. Pro- sive of Sirenia and Cetacea. Sanseido Company, Tokyo (in ceedings of the Royal Society of London. Series B: Biological Japanese). Sciences, 269: 801–807. Kuroda S. (1984). Morphological characters and evolution of