Phenotypic and genotypic variation and structure of 29 sour cherry (P. cerasus) and duke cherry (P. x gondouinii) genotypes from different regions of Iran were identified using random amplified polymorphic DNA (RAPD) markers and morphological characters. Furthermore, one Prunus mahalebgenotype was used as an outgroup for molecular analysis. For morphological analysis, 23 variables were recorded to detect similarities between and among studied sour and duke cherries. Most studied characteristics were showing a high degree of variability. Principal component analysis showed that the first three components explained a total of 73.87 % of the whole phenotypic variability. Based on the morphological cluster analysis, studied sour and duke cherry genotypes were placed into three main clusters. The first main cluster included 16 sour cherry genotypes. The second main cluster contained all duke cherry genotypes and eight sour cherry genotypes, while, only one sour cherry genotype was placed in third main cluster. For RAPD analysis, 17 primers generated a total of 233 discernible and reproducible bands across genotypes analyzed, out of which 214 (91.51 %) were polymorphic with varied band size from 300 to 3000 bp. According to the similarity matrix, the lowest similarity was obtained between P. mahaleb, as an outgroup, and sour cherry. Dendrogram based on molecular data separated genotypes according to their species and geographic origin. Low correlation was observed between the similarity matrices obtained based on morphological and RAPD data. The information obtained here could be valuable for devising strategies for conservation of Iranian sour and duke cherries.
