Rediscovering the value of images in supporting personal identification of missing migrants.

The crisis concerning the identification of dead migrants is forcing forensic scientists to deal with the unavailability of classical antemortem (AM) data such as radiological information and DNA. However many times there is abundant AM photographic data of which the availability and quality is underestimated. Independently of when facial image comparison can be used as a valid identification method, description of the unexpected quality and quantity of images in the scenario of unidentified migrant decedents is given in this observational study. It focuses on the overall characteristics of the AM images available for a group of 74 missing migrants related to the disaster of October 3rd, 2013. 394 images were collected: an average of 6 for each missing person. 18% are original pictures, 47.5% scans of original images, 9.9% scans of printed images while 9.4% downloaded from social networks such as Facebook. 42.7% of images resulted in focus. Among all the images, about the 10% has been used in the personal identification process to ascertain the identity of 22 persons, because of the visibility of identifying details (facial, dental and body features). The importance of such a finding should not be diminished by its simplicity. Today, those involved in collecting AM information from relatives of dead migrants are underestimating the possible presence of images, even in sub-Saharan populations, where, contrary to belief and data, technology, though simple, is available and images may turn out to be abundant and useful (and at times the only resort) for identification.

Preliminary study on sexual dimorphism of metric traits of cranium and mandible in a modern Italian skeletal population and review of population literature.

This study aims to assess the sexual dimorphism of skull in a modern skeletal collection through a direct anthropometric methodology by using a logistic regression analysis as statistical approach in order to provide specific regression formulae for the Italian population. Thirtyfour measurements (24 from cranium, 10 from mandible; 33 linear distances, one angle) were taken on 80 individuals (40 males, 40 females). A stepwise discriminant function analysis selected the combination of variables which best discriminated between sexes, and a cross-validation assessed the accuracy rate of the original sample. Most distances resulted longer in males than in females, but differences were statistically significant only for 12 and 6 out of respectively 24 cranium and 10 mandibular distances. Five combined cranium features returned a prediction accuracy of 88.6% (y = -81.01 + 0.14* Maximum cranial length + 0.16* Basion-bregma height -0.05* Cranial base length +0.20* Bizygomatic breadth +0.31* Nasal height; adjusted R2 = 47.0%). Nine mandibular features were required to reach 74.7% accuracy (y = -93.08 + 0.19* Bigonial width +0.14* Bicondylar breadth +0.11* Mandibular length +0.18* Height of the mandibular body at mental foramen +0.21* Chin height -0.09* Mandibular angle -0.06* Minimum ramus breadth -0.05* Maximum ramus height -0.01* Maximum ramus breadth; adjusted R2 = 23.92%). Current results and existing literature suggested that only few cranial measurements can be used for diagnosis of sex: the same variables showed similar accuracy in different ethnic contexts. In conclusion, some aspects of sexual dimorphism of skull seem to be independent from ancestry.

Population specific data improves Fordisc®'s performance in Italians.

Using discriminant functions based on morphometric data is one of the most approved methods for sex and ancestry estimation on skeletons. Nevertheless, population data from various parts of the world is needed in order to reliably classify an individual into a group. Due to population variation even sex estimation is biased when there is a lack of adequate data. Software that computes discriminant functions based on morphometric data is Fordisc®. Unfortunately, the above mentioned effects reduce its applicability in countries other than the US. For improvement of this situation data collection is currently performed extensively. The present paper shows a comparison of an Italian sample from the identified modern skeletal collection of CAL (Collezione Antropologica Labanof) [1] (Cattaneo, 2018) at the Institute of Legal Medicine Milan, Italy with a Euro-American sample from the Forensic Data Bank at the University of Tennessee Knoxville, USA. Fordisc® 3.1 was used to study population differences and sexual dimorphism. The analyzes were performed on a selection of 19 highly influential measurements that are present in most individuals of both groups respectively. Italian crania show in relation to Euro-Americans wider and lower vaults with shorter cranial bases and wider faces. The degree of sexual dimorphism is similar in both groups. Yet there is a shift in the absolute value range for males and females that biases sex estimation by almost 25% when an individual is classified on the respective other discriminant function. Our results provide explanations for Fordisc®'s unsatisfying performance on non-US individuals. At the same time they show that significant improvement is achieved by adding more population samples to its dataset.