prilagođeno pretraživanje po punom tekstu

ŠUMARSKI LIST 7-8/2022 str. 38     <-- 38 -->        PDF

silver fir needle traits were estimated by the analysis of variance (ANOVA, procedure PROC GLM in SAS software). Testing the significance of differences between populations (as fixed factor) and trees nested in population (as a random factor) was performed by F-test, using RANDOM options in PROC GLM procedure. The PROC CORR option in the SAS software was used to estimate Pearson’s correlation coefficient between morphological needle traits and climatic factors. The Principal Component Analysis (PCA), a multivariate method, was applied for the analysis of population variability. In order to examine the patterns of population variability in relation to the morphological needle traits and climatic factors PCA based on the correlation matrix (Pearson’s method) was performed. We used scaterplot graph to visually present results both on populations, morphological needle traits and climatic factors. The Microsoft Excel XLSTAT add-in software package Agglomerative Hierarchical Clustering (AHC) (Ward’s method) was performed on standardized mean values of needle traits using Euclidean distance, which provided the optimal classification of the analyzed populations into homogeneous groups, so-called clusters. The cluster analysis was visually presented by a dendogram.
Statistical data analyses were performed using the appropriate procedures in the software package SAS 9.1.3. Graphic presentations of the obtained results were performed by XLSTAT in Microsoft Excel add-in software package.
Mean values of the morphological needle traits varied between populations. The needle length varied from 2.08 to