JWSS - Journal of Water and Soil Science

---

The precision of evaluation for agronomic traits is very important in the field experiments to determine the genetic potential of genotypes or effect of treatments. The precision of the experiment mainly depends on the number of replications. The objective of this study was to investigate the effect of number of replications on the precision of genotype evaluation, estimation of variance components and on the heritability for agronomic traits in flax (Linum usitatissimum L.). In this study, nine genotypes of flax were evaluated for different traits, using a randomized complete block design with 8 replications. The results showed that the reduction of error variance due to increasing the number of replications mainly depended upon the type of the trait. For evaluation and estimation of heritability for number of seedlings /m² and plant height, 3 and 2 replications were suitable, respectively. The results also showed that using more than 2 replications had no considerable effect on the prescision of evaluation, nor on the increase of heritability for days to maturity. Based upon the results, for evaluation of the number of capsules per plant, 3 or 4, and for seeds per capsules, 4 replications can be suggested. For seed yield that is the most important economic trait in flax, a relatively constant estimated mean and broad-sense heritability were obtained with 4 to 8 replicationshowever, using 2 or 3 replications over-estimated the mean of the triat (approximately 11%) and under-estimated the broad-sense heritability about 27% and 10%, respectively. With 2 replications, the estimated mean and broad-sense heritability for seed yield were 2017.6 kg/ha and 72.1%, respectively. Gain from selection for seed yield based on evaluation of genotypes with 2, 3 and 8 replications was estimated as 787, 796 and 798 kg/ha, respectively. Therefore, it seems that 2 replications can also be used for seed yield evaluation of genotypes in breeding programs of flax.