Fruit remains such as cherries, plums, sloes, peaches from archaeological waterlogged contexts as wells, latrines and lake dwellings are frequently recovered. Although, commonly the discrimination among Prunus species is based on visual evaluations of some traditional morphological characters of the fruit stones, currently, computer vision and image analysis techniques represent more accurate, reliable and repeatable alternative to distinguish wild species from cultivated ones (Bacchetta et al., 2008; Venora et al., 2009). In this work are presented the results obtained from discrimination by image analysis technique of the Prunus L. archaeo-endocarps found from two archaeological sites located in the South-West of Sardinia, dating from the Bronze Age and Punic period. All endocarps due to the waterlogged sediment was in excellent condition that has preserved the biological material. One hundred and six morphometric features of the endocarps, including the Elliptic Fourier Descriptors (EFDs), were measured by the open source software ImageJ v.49. Archaeo-seeds were compared with 27 modern populations of Prunus spinosa L. and 22 varieties of Prunus domestica L. collected from the field catalog of CNR-ISPA (Nuraxinieddu, OR) in Sardinia. The achieved results were used to implement a database of morphometric variables that was applied to carry out a statistical classifier, based on the stepwise Linear Discriminant Analysis method, able to compare and distinguish the unknown archaeo-seeds with modern ones. The results have enabled correctly discriminate all archaeo-seeds, this suggest that for the genus Prunus multivariate statistical technique, based on the morphometric parameters, is a useful tool to discriminate different group of species.
Prunus L. endocarps from two archaeological sites in Sardinia: characterization by image analysis
BACCHETTA, GIANLUIGI
2014-01-01
Abstract
Fruit remains such as cherries, plums, sloes, peaches from archaeological waterlogged contexts as wells, latrines and lake dwellings are frequently recovered. Although, commonly the discrimination among Prunus species is based on visual evaluations of some traditional morphological characters of the fruit stones, currently, computer vision and image analysis techniques represent more accurate, reliable and repeatable alternative to distinguish wild species from cultivated ones (Bacchetta et al., 2008; Venora et al., 2009). In this work are presented the results obtained from discrimination by image analysis technique of the Prunus L. archaeo-endocarps found from two archaeological sites located in the South-West of Sardinia, dating from the Bronze Age and Punic period. All endocarps due to the waterlogged sediment was in excellent condition that has preserved the biological material. One hundred and six morphometric features of the endocarps, including the Elliptic Fourier Descriptors (EFDs), were measured by the open source software ImageJ v.49. Archaeo-seeds were compared with 27 modern populations of Prunus spinosa L. and 22 varieties of Prunus domestica L. collected from the field catalog of CNR-ISPA (Nuraxinieddu, OR) in Sardinia. The achieved results were used to implement a database of morphometric variables that was applied to carry out a statistical classifier, based on the stepwise Linear Discriminant Analysis method, able to compare and distinguish the unknown archaeo-seeds with modern ones. The results have enabled correctly discriminate all archaeo-seeds, this suggest that for the genus Prunus multivariate statistical technique, based on the morphometric parameters, is a useful tool to discriminate different group of species.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
