We can judge whether it is human or not by glancing at its morphology.  This is probably because there are limits and some regularity in the morphological variations of the animal called human being.  The ultimate aim of the present author is to understand the mechanism of evolutionary formation of our body structure.  As a step to achieve the ultimate aim, it is attempted in the present study to confirm the limits and regularity of among-group variations in the craniofacial morphology of Homo sapiens sapiens, and, if possible, to determine some of the causes for the regularity of among-group variations, i.e., for the among-group covariations between craniofacial measurements.

    As regards the within-group covariations of morphological traits, a lot of studies have been carried out.  For example, Howells (1957, 1972, 1973), Kanda and Kurisu (1967, 1968), Kanda (1968), and Brown (1973), using multivariate statistical methods, found that there are some common factors controlling the craniofacial morphology.  Furthermore, Mizoguchi (1992, 1994, 1995b, 1996, 1997, 1998a, d, 1999, 2000a, 2001, 2002, 2003a, b, 2004a, b, 2005, 2007a, b, 2008, 2009, 2013a) carried out a series of principal component analyses (PCAs) of within-group correlations between cranial and postcranial measurements mainly to elucidate the causes of brachycephalization on the premise that population differences are extensions of individual differences, as stated by Howells (1973).  As a result, he found several common factors suggesting that, while cranial breadth has no consistent associations with any postcranial measurements, cranial length is significantly associated with many postcranial measurements, such as vertebral body size, costal chord, pelvic widths, and limb bone lengths and thicknesses; and considered that the variation in cranial length may, in part, be related to the degree of development of skeletal muscles or body size and, besides, that the form of the maternal pelvic inlet may be another important determinant of neurocranial form.

    Common factors extracted from PCAs or factor analyses are, however, usually interpreted as, for example, a cranial length factor, a lower face factor, etc. according to the properties of the original variables which are strongly correlated with the common factors in question.  Such analyses do not inform us whether the common factors extracted are pleiotropic genes, common environmental factors, or a composite of them unless any candidates of causes for the variations of biological characters under consideration are included in the data sets to be analyzed.

    Recently, however, from another angle, molecular biology made it possible for us to know the correspondence of some morphological characters with gene loci on chromosomes.  Dorus et al. (2004) compiled a list of 27 genes demonstrated to play important roles in the nervous system including the brain, and discussed the evolution of the human brain.  Coussens and van Daal (2005) found that a single-nucleotide polymorphism (htSNP g.8592931G->C) in the gene FGFR1 (fibroblast growth factor receptor 1) had a significant negative correlation with the cephalic index for all of four populations, i.e., so-called Caucasian, Asian, Australian Aboriginal, and African American populations.  Evans et al. (2005) maintain that the gene Microcephalin (MCPH1) regulates brain size and has evolved under strong positive selection in the human evolutionary lineage.  Mekel-Bobrov et al. (2005) state that the gene ASPM (abnormal spindle-like microcephaly associated) is a specific regulator of brain size, and that its evolution was also driven by strong positive selection in the lineage leading to Homo sapiens.  Liu et al. (2012), using almost ten thousand individuals of European descent, identified five independent genetic loci (at 1p36.23-p33, 2q35, 3q28, 5q35.1, and 10q24.3) associated with different facial phenotypes.  The candidate genes involved with these five loci are PRDM16 (PR domain containing 16), PAX3 (paired box 3), TP63 (tumor protein p63), C5orf50 (chromosome 5 open reading frame 50), COL17A1 (collagen, type XVII, alpha 1).  Liu et al. contend that their finding at PAX3 influencing the position of the nasion replicates a genome-wide association study of facial features independently performed by Paternoster and others in 2012.  Shaffer et al. (2016) observed genome-wide significant associations for cranial base width at 14q21.1 and 20q12, for intercanthal width at 1p13.3 and Xq13.2, for nasal width at 20p11.22, for nasal ala length at 14q11.2, and for upper facial depth at 11q22.1 on the basis of European data.  They also tested genotype-phenotype associations reported in two previous genome-wide studies and found evidence of replication for nasal ala length and SNPs in CACNA2D3 and PRDM16.  Roosenboom et al. (2018) performed a genome-wide association study on three vault measures (maximum cranial width, maximum cranial length, and cephalic index) in a sample of 4419 healthy individuals of European ancestry, and observed significant associations at two loci: 15p11.2 for maximum cranial width and 17q11.2 for maximum cranial length.

    In the near future, genome-wide association studies will identify all loci for morphological characters.  And the correspondence of all genes to their functions will also be clarified in molecular biology or related fields.  However, there remain other questions to be answered, as was pointed out by Mizoguchi (2000b, 2006, 2013b): when, where, and how did such genes appear and become fixed in ancestral populations?  It would be impossible to determine the causes and mechanisms of their appearance and fixation if we only explore genes in living human populations or if we analyze only within-population variations of genes or morphological characters.  To elucidate the causes and mechanisms, we must collect data not only on morphological characters or their associated genes in ancient populations but also on ancient environments where morphological characters first came into existence.  At present, however, we do not have sufficient paleoecological data for this purpose.  It is our task for the future.

    Hence, the use of data on environmental factors in the present day may be recognized as the next best alternative to search for the causes for the appearance of morphological characters in human evolutionary processes.  In practice, some researchers have already examined ecological correlations (Yasuda, 1969; a.k.a. among-group or inter-population correlations) between morphological characters and climatic factors in modern times.  For example, Beals (1972), Guglielmino-Matessi et al. (1979), Beals et al. (1983, 1984), Mizoguchi (1985), and Kouchi (1986) show that cephalic index is higher in colder regions or in higher latitudes.  Weiner (1954), Wolpoff (1968), Yamaguchi (1970), Carey and Steegmann (1981), and Mizoguchi (1985) state that nasal breadth is smaller, or nasal index is lower, in higher latitudes or colder and drier regions.  Crognier’s (1979, 1981), using European, North African and Near/Middle Eastern samples, quantitatively showed that not only head and face dimensions but also body size were significantly correlated with temperature and precipitation.

    Mizoguchi (1998b, c) also performed among-group analyses of craniofacial measurements on the basis of 308 male and 200 female Asian samples from the past 10000 years, and found that, while cranial breadth, bizygomatic breadth, upper facial height, and nasal height always varied in parallel with one another, cranial length and nasal breadth varied independently of each other and of the above four measurements.  Later, Mizoguchi (2007a) preliminarily estimated ecological correlations between neurocranial and limb bone measurements simply using Spearman’s rank correlation coefficient on the basis of 24 male and 23 female samples from prehistoric, protohistoric, medieval, early modern, and modern populations in Japan.  The results pointed to significant associations between cranial length and the thickness measurements of the radius, ulna, femur, and tibia in both males and females.

    For the above associations, various causes can be considered.  Namely, pleiotropic genes, linkage of genes, the state of two characters being elements in the same ontogenetic process, physiological cycle or biomechanical causation, etc. may cause both intra- and inter-group correlations between characters.  Further, an ecological correlation between a character and an environmental factor or between two genetically/ontogenetically independent characters may result from one or more of three basic evolutionary causes, i.e., adaptation to local environments through natural selection, random genetic drift, and gene flow (i.e., migration and/or hybridization with other populations), as suggested by many authors (e.g., Stern, 1960; Dobzhansky, 1963; Mettler and Gregg, 1969; Harrison at al., 1977; Molnar, 1992; Frisancho, 1993; Marks, 1995; Mizoguchi, 2013b).

    Mizoguchi (2014), keeping differential contributions of such possible causes in mind, performed a preliminary PCA of among-group correlations between three cranial and four postcranial measurements as well as latitude and chronological age.  But the data used there are of only 14 male samples from the Japanese archipelago of the Jomon period to modern times.  The present study is an extended version of it.  After many more data were collected, the same analyses were performed here to achieve the aim mentioned at the beginning.