Moscow University
Anthropology
Bulletin
Anthropology
Bulletin
Lomonosov Moscow State University, Anuchin Research Institute and Museum of Anthropology, Mokhovaya st., 11, Moscow, 125009, Russia
Evteev A.A., ORCID ID: 0000-0002-6254-1203, e-mail: evteandr@gmail.com
Objective. Better knowledge of correlation structure of craniometric variables describing middle facial skeleton could potentially help to achieve parsimony in the set of measurements commonly used when studying this part of the skull. Thus by analyzing that structure we were aimed to: – find potentially informative variables among those that are rarely used at present – figure out what variables are redundant due to their high correlation with the others. Methods. The sample comprises 362 male skulls representing many craniologically divergent populations from around the world. 44 additional variables were measured from each skull by the author according to previously published protocol [Evteev, 2010] as well as 26 variables of «standard» craniometric blank [Alexeev, Debets, 1964; Howells, 1989]. Two exploratory statistical methods we applied: PCA based on correlation matrix and non-metric MDS based on matrix of correlation distances among the variables. Results. On the first stage of the analysis it has been shown that numerous measurements of mid-facial protrusion are strongly correlated among themselves while being distinct from all other variables. The same is true for measurements of breadth of either frontal process of the maxilla or the nasal bones. The analyses done after exclusion of those distinctive measurements revealed much less «structured» picture of inter-variable correlations, in other words there were no clusters of strongly intercorrelated measurements apparently falling out from the general scatter. Nevertheless there were also several interesting tendencies to point out: – singularity of variables describing orbital contour including height of the frontal process of the maxilla and measurements of the zygomatic process of the same bone – sagittal length of the maxilla is not fully correlated with neither its height nor its width – height and width of the choanae as well as breadth of the palate are weakly correlated with most of other variables – some measurements of the lower part of the piriform aperture possess their own tendencies of variation independent from total height or breadth of the aperture – lower and upper parts of the nasal bones seem to be relatively «independent» from each other – upper facial height (48 Mart.) is strongly correlated with alveolar process height which is known to be strongly epigenetically influenced. The last analysis which included a number of neurocranial measurements confirmed well-established “autonomy” of two main cranial regions. But two variables, biauricular breadth and maximum cranial breadth were found to be much stronger correlated with the maxilla’s measurements than all other neurocranial dimensions. Conclusions. Several practical recommendations based on the results of the analyses conducted are formulated. There are several “positive” points stating which from rarely used variables could have potentially been useful in providing additional data about mid-facial form if included in the craniometric program, and also «negative» ones – what measurements can be excluded from the program without losing important information.
craniology; craniometry; viscerocranium; correlation structure
Цит.: Evteev A.A. The analysis of the correlation structure of mid-facial dimensions as the basis for the optimization of the craniometric programm // Moscow University Anthropology Bulletin (Vestnik Moskovskogo Universiteta. Seria XXIII. Antropologia), 2014; 4/2014; с. 18-30
Download text