Articles
HAZELNUT CULTIVAR IDENTIFICATION WITH LEAF MORPHOMETRIC ANALYSIS: PRELIMINARY RESULTS
Article number
845_34
Pages
245 – 248
Language
English
Abstract
The variety identification in hazelnut is an important aspect in the nursery and in the orchard.
Methods based on molecular biology are expensive, complex and often ineffective.
Plant taxonomists recognized the importance of leaf features for identifying taxa a long time ago.
In this preliminary work, a comparison of leaf morphology for evaluating differences between three cultivars was conducted through the geometric morphometric analysis.
Twenty-four leaves collected from trees at the field station in Fiorano-Rome for three commercial cultivars (Butler, Ennis, Tonda Gentile Romana) of hazelnut (Corylus avellana, L. 1753) were analyzed via geometric morphometric tools, after their image scanning (300 d.p.i.; 24 bit color). This method allows the visualization and the quantification of morphological differences by representing the deformations of morphometric points (landmarks) in a coordinate space.
Twenty-eight landmarks along the margins and at the basis of the veins on the inferior side of each leafs scanned image were digitized.
A Canonical Variates Analysis (CVA) was performed on the obtained weight matrix (W) in order to discriminate among cultivars.
The three cultivars were discriminated with an efficiency of 100%. Further trials on many more samplings and cultivars will be useful to test, implement and give more reliable results.
The preliminary results obtained indicated the technical feasibility of the proposed method.
Methods based on molecular biology are expensive, complex and often ineffective.
Plant taxonomists recognized the importance of leaf features for identifying taxa a long time ago.
In this preliminary work, a comparison of leaf morphology for evaluating differences between three cultivars was conducted through the geometric morphometric analysis.
Twenty-four leaves collected from trees at the field station in Fiorano-Rome for three commercial cultivars (Butler, Ennis, Tonda Gentile Romana) of hazelnut (Corylus avellana, L. 1753) were analyzed via geometric morphometric tools, after their image scanning (300 d.p.i.; 24 bit color). This method allows the visualization and the quantification of morphological differences by representing the deformations of morphometric points (landmarks) in a coordinate space.
Twenty-eight landmarks along the margins and at the basis of the veins on the inferior side of each leafs scanned image were digitized.
A Canonical Variates Analysis (CVA) was performed on the obtained weight matrix (W) in order to discriminate among cultivars.
The three cultivars were discriminated with an efficiency of 100%. Further trials on many more samplings and cultivars will be useful to test, implement and give more reliable results.
The preliminary results obtained indicated the technical feasibility of the proposed method.
Publication
Authors
C. Costa, G. Paglia, F.R. De Salvador, D. Lolletti, V. Rimatori, P. Menesatti
Keywords
hazelnut, shape, geometric morphometric, cultivar identification, CVA
Online Articles (125)
