Labokas, J. 2013 Quantitative Evaluation of Leaflet Shape in Fragaria vesca and F. viridis Using Image Analysis and Elliptic Fourier Descriptors. Baltic Forestry 19(1): 22 - 30

The aim of the study was to emphasize the taxonomic utility of the central leaflet shape variation in Fragaria vesca and F. viridis considering the factors that may influence it. The image analysis and principal component analysis of elliptic Fourier descriptors of leaflet contours as well as leaflet area, perimeter and shape index, all derived from image outline, were employed. Simultaneously, the identification of diversity patterns of the species was pursued. 100 leaflets from each of six populations of each species were sampled across Lithuania. For F. viridis a significantly larger leaflet area and more circular leaflet shape, as estimated by shape index, are typical. The analysis of variance of principal component (PC) scores showed that for both species symmetrical shape variation is much more typical than asymmetrical. If estimated by the first five PC, the symmetrical variation accounts for about 60 % of total variation in F. vesca and 67 % in F. viridis, while the asymmetrical variation is up to 22 % and 18 %, respectively. Hierarchical cluster analysis revealed that PC7 and PC8 are good indicators of species-wise grouping of populations. Cluster analysis of shape index provided informative clustering results of populations within and between species.

Key words: Fragaria vesca, Fragaria viridis, population, shape index, leaflet shape, elliptic Fourier descriptors, image analysis.