---

In order to select the best parents for crossings, plant breeders seek varieties or genotypes with highest genetic dissimilarities. This can be achieved by measuring the similarities among genotypes, using multivariate analysis methods such as cluster analysis. This study aimed to group 50 aerobic and lowland rice genotypes based on biochemical characteristics including Iron, Zinc, Manganese and protein, and their linked DNA markers. According to the cluster analysis results using Ward method, the genotypes were assigned to four groups. The third group, as the smallest group including three genotypes (IR82635-B-B-82-2, Caiapo, and Gohar), had the highest value for these micronutrients. Their mean value for Iron, Zinc, Manganese, and protein were 32.39, 34.15, 25.66 mg/kg and 6.71%, respectively. Also, all genotypes were classified into two main groups based on microsatellite markers information, that according to QTL mapping studies these markers were identified as linked to elements. So, the most of non-local genotypes and aerobic rice cultivars were assigned in a separate group. The correlation between Euclidean distance of elements and protein matrix and genetic similarity matrix (Nie) using Mental correlation test was estimated significant (p<0.01) that can be evidence of a genetic relationship between the SSR markers and genome controlling regions of elements in this population.

---

Various strategies have been developed to reduce water consumption in rice cultivation, including rice cultivation in water-saturated soil, drying and wetting of the soil (periodic irrigation), and aerobic rice cultivation. In this study, in order to identify and grouping rice cultivars suitable for aerobic cultivation, 34 cultivars of native and improved rice cultivars were selected and evaluated. Factor analysis based on principal components analysis based on all traits showed that 6 factors were extracted with special values of more than one. These 6 factors, after the Varimax rotation, justify the 6 main and independent factors, justifying 78.08% of the total variation. In the first harvest, cluster analysis using WARD method, cultivars were grouped into 4 groups, which cluster 3 being the largest cluster. Discriminant analysis of first harvest, showed that all cultivars were grouped correctly. In canonical discriminant the first two canonical functions with eigenvalues greater than one explained a total of 88.9% of the variance. In cluster analysis based on the average of the data obtained from Raton product, genotypes were divided into two different groups. The genotypes of the first group were higher than the genotypes of the second group in terms of yield and yield components and were introduced as superior genotypes. Heritability with genetic improvement for leaf area duration, number of seedling (grown plants), number of spikes per unit area, number of full seeds and degree of maturity showed that these traits can be transferred to hybrid offspring and selections based on these traits are effective.