Path analysis of flower yield and its components in Iranian Rosa damascena genotypes

Flower yield in Rosa damascena is a complex trait that is affected by many traits and environmental factors. Therefore, studying the relationships among traits can be an essential step in selection to improve flower yield. Path analysis by identifying the main components of yield with a high direct effect plays an important role in the selection success, especially indirect selections through yield components. Due to the lack of resources in this regard, the present research aims to identify and determine the effective traits in flower yield, estimate the relationships among traits, and analyze the causality to identify their direct and indirect effects on flower yield via path analysis. 

The germplasm used in this research included 49 genotypes of Rosa damascena Mill.) belonging to different regions of Iran. One-year saplings of the genotypes were planted in April 2004 in the Zaleh research station, Sanandaj (Kurdistan province, Iran), based on the randomized complete block design (RCBD) with three replications and three plants in each plot with three meters of planting distance. Irrigation of the field was carried out in drip form for 15 days. The investigated morphological traits (plant height, number of flowers, etc.) in 2008 were measured on a 5-year old plant, and for the phenological traits (days until leaf opening, flower opening, etc.) the average of three years (2006 to 2008) was used. Data variance analysis was done based on RCBD design. Phenotypic correlation among traits was estimated, and path analysis was done in three steps of estimation of direct, indirect and residual effects of traits on flower yield.

The results showed significant differences (p≤0.01) among the genotypes for studying traits. The correlations of phonological traits with flower yield (FY) were significantly negative, while its relationships with morphological traits were significantly positive. The results of path analysis showed that the number of flowers per plant (FN) and fresh flower weight (with the direct effects of P=0.97** and P=0.33**, respectively, and residual effects of R=0.195) were the main components of flower yield. The total green surface of the plant (TGS) and plant crown diameter (with the direct effects of P=0.56** and P=0.28*, respectively, and R=0.395) were the main components of FN. Also, the number of leaves per plant and average leaf area (with the direct effects of P=1.02** and P=0.085*, respectively, and R=0.237) were the main components of TGS. 

Thus, FN and FFW can be used in indirect selections for FY improvement in Rosa damascena. The other studied traits affected FY indirectly via FN mainly (phonological traits negatively and morphological traits positively). Although the correlation of the height and crown diameter of the plant with FY was positive and very significant, the direct effects of the supplementary stages and the causality analysis sequence showed that the crown diameter (horizontal growth) is more effective than its height (vertical growth) in the reproductive processes of the rose. In other words, the genotypes with higher flower yield usually rise up to a certain height and spread leaves after that.