DIGITALNA ARHIVA ŠUMARSKOG LISTA
prilagođeno pretraživanje po punom tekstu



ŠUMARSKI LIST 1-2/2018 str. 27     <-- 27 -->        PDF

with the total CV of 24.27%. As opposed to that, the lowest value of the coefficient of variation was found for the angle closed by the main leaf vein and the line defined by the leaf blade base and a point on the leaf margin, at 10% (LA1 = 10.46%) and 25% (LA2 = 7.82%) of leaf blade length.
The results of the hierarchical analysis of variance (ANOVA) for each characteristic are shown in Table 1. The studied regions differ significantly for all studied variables. Likewise, statistically significant differences among trees within populations were observed. The AMOVA analysis showed that most of the morphological diversity was attributable to the differences between regions, confirming the geographical structuring of populations. However, a highly significant percentage of variation was explained by the differences among individuals within populations for the variables: leaf blade length, measured from the leaf base to the point of maximum leaf width (PMLW); leaf blade width at 90% (LW2); petiole length (PL); and leaf angles LA1 and LA2. Among populations within regions, there were significant differences in four of the nine analysed traits (LA, MLW, LW1, PL), where only a small proportion of the total variation was explained.
As expected, strong correlations between leaf morphological traits were observed. Almost all measured leaf traits correlated with each other at a statistically significant level. The highest correlations, considering all populations, were found between leaf blade area (LA); leaf blade length (LL); maximum leaf blade width (MLW); and leaf blade width at 50% of leaf blade length (LW1). We also observed strong and significant correlations between the angles closed by the main leaf vein and the lines defined by leaf blade base and points on the leaf margin, at 10 % (LA1) and 25% (LA2) of leaf blade length. Non-significant correlations were found between petiole length (PL) and the above-mentioned leaf traits (LA1 and LA2).
Using Spearman’s correlation coefficient (Table 2), a highly positive correlation was found between geographical longitude and the six measured variables (LA, LL, MLW, LW1, LA1, LA2). On the other hand, all measured variables were highly negatively correlated with the altitude and total precipitation of the warmest quarter of the year. Likewise, almost all leaf traits were significantly negatively correlated with the distance-to-water, annual mean precipitation, precipitation of the coldest quarter and isothermality. By contrast, almost all leaf traits were positively correlated with the annual mean temperature, maximum temperature of the warmest month and precipitation seasonality.
The Mantel test identified significant correlations between the morphological, geographical, and environmental distance matrices. Correlations were higher between morphological and geographical distance matrices (r = 0.91, p = 0.0018), and slightly smaller but still very strong between the morphological and environmental distance matrices (r = 0.80, p = 0.0029).
The structure of the seven grey alder populations was inferred by the K-means clustering method. The most probable division was detected at K = 2, and the estimated population structure is shown in Figure 1. If the proportion of a certain population was equal to or higher than 0.75, it was assumed that the population belonged to one cluster, and if it was lower than 0.75, it was assumed that the population had a mixed origin. The populations from the Alpine-Dinaric region grouped together into cluster A, and the populations from continental Croatia grouped together into cluster B. Only one population, P05 (Varaždin), was of mixed origin, with the dominant proportion from cluster B. The results obtained with the K-means clustering method were congruent with the hierarchical tree (Figure 4C), where the unweighted pair-group method with arithmetic mean (UPGMA) was used to join the clusters.