Dissertation committee:
Κωνσταντίνος Μωραΐτης, Αν. Καθηγητής, Ιατρική, ΕΚΠΑ
Κωνσταντίνος Ηλιόπουλος, Επ. Καθηγητής, Department of Natural Sciences and Psychology, Liverpool John Moores University
Ευστράτιος Βαλάκος, Καθηγητής, Βιολογία, ΕΚΠΑ
Χάιδω Σπηλιοπούλου, Καθηγήτρια, Ιατρική, ΕΚΠΑ
Δημήτριος Βλαχοδημητρόπουλος, Αν. Καθηγητής, Ιατρική, ΕΚΠΑ
Χριστίνα Παπαγεωργοπούλου, Αν. Καθηγήτρια, Ιστορία και Εθνολογία, Δημοκρίτειο Πανεπιστήμιο Θράκης
Μαρία Πιάγκου, Επ. Καθηγήτρια, Ιατρική, ΕΚΠΑ
Summary:
In commingled contexts, the accurate sorting of human skeletal remains constitutes the basis for further anthropological examination, including the identification of the individual’s basic biological characteristics. Particularly, the morphology of certain bone elements (i.e., scapula, humerus, radius, ulna, os coxa, femur, tibia, talus, and calcaneus) is widely used by forensic anthropologists for accurate estimation of sex, age, body height, and body mass. In this doctoral thesis, a total of 23 standard anthropological measurements were obtained in a total of 214 adult individuals (114 males, 100 females) originating from the "Athens Collection". The Collection represents a sample of late 20th century individuals who lived in the broader region of Athens, Greece. Simple and multiple regression analysis were used for developing a total of 22 mathematical equations (16 simple, 6 multiple) for sorting commingled human bone remains based on the dimensions of adjoining articular surfaces. The results of these analyses revealed very strong and positive correlations (r=0,69 – 0,95, p-value <0.05), relatively high coefficients of determination (r2=0.47 – 0.91) and low standard error of the estimate (SEE=0.88–2.14). Bilateral asymmetry and sex did not have an effect on the accuracy of the functions developed.
For future application of the methodology introduced in this doctoral thesis, two steps are recommended. The first step requires the employment of the metric method developed. Particularly, linear measurements of a skeletal element or multiple skeletal elements are used to predict a dimension of another skeletal element. This leads to a small number of possible matches for each bone specimen. The second step involves a systematic morphoscopic examination of these matches, in order to determine the best fit.
For further demonstrating the applicability of the introduced methodology, consequent blind tests were performed in random samples of 8 to 20 individuals which were not part of the material used to develop the regression equations. The
results showed that the percentage of correct classification was 100% for most joints examined. Further research is necessary for improving the classification rates for certain joints (shoulder, ankle, subtalar) on the basis of alternative approaches, such as the application of three-dimensional geometric morphometrics. Furthermore, future forensic research on commingled bone assemblages would benefit from further expanding these sorting methods to other joints of the human skeleton, aiming to gradually provide a reliable methodological toolkit for reconstructing the biological characteristics of each individual skeleton.