Morphological Parameters in Grape Bunches and Berries Evaluation through Multivariate Analysis – Case Study on the 'Feteasca Regala' Cultivar

Authors

  • Alin DOBREI Author
  • Florin SALA Author

DOI:

https://doi.org/10.58509/0jpjmz10

Keywords:

cluster structure, Correspondence Analysis, factor loading, grape berry categories, Multivariate Analysis

Abstract

Morphological parameters of grape bunches and grape berries in bunches are important indicators for the yield and quality of grapes and wine products. The study analyzed structural elements of grape bunches and berries, the 'Feteasca Regala' cultivar. The grape samples came from the Recas Viticultural Center. Several parameters were determined: grape bunch weight (GBW, g); four categories of grape berry, normal grape berries (C1), mean/small grape berries (C2), small grape berries (C3), spoiled grape berries (C4) – for each category the number (no), and the weight (w, g). The total  berry number of grape (TBnG) was calculated. The rachis weight (RW, g), and must volume (MV, ml) were determined. According to the coefficient of variation, all parameters showed low variability. Very strong, positive correlation was recorded between C4-no and C4-w (r = 0.934, ***), and very strong negative correlation was recorded between C1-w and C4-w (r = -0.922, ***). Strong, positive correlation was recorded between TBnG and C1-no (r = 0.816, ***), between TBnG and C2-no (r = 0.834, ***), between C2-no and C2-w (r = 0.897, ***). These correlation values ​​showed the high importance of the two categories of grape berries (C1 and C2) in the formation of the TBnG parameter. According to PCA, five parameters were loaded in Factor 1, with very strong action (C1-w, C4-w, C4-no), with strong action (C3-w), and with moderate action (C2-no). In Factor 2, two parameters with strong action (TBnG, C1-no) were grouped. In Factor 3, the RW parameter with strong action was positioned. Correspondence Analysis showed the score values ​​of each parameter in relation to the working Axes, and the distribution of the parameters based on the recorded score.

Downloads

Published

06.08.2025