Investigating relationships between yield and yield components in promising cotton genotypes (Gossypiume hirsutum L.)

Background and Objective: Cotton yield and most important traits in terms of economic and breeding values are polygenic traits, and direct selection for them has not been very successful. This study was conducted to identify the relationships between yield and its related traits and to determine the most important traits affecting cotton lint-yield.
Materials and Methods: an experiment including nine genotypes with two commercial varieties of Fars province (Bakhtegan and Golestan) as controls, in a randomized complete blocks design with four replications, was carried out at Darab Agricultural Research Station during two years of 2017 and 2018. In this study, the traits of plant height, monopodial length, monopodial number, sympodial length, sympodial number, number of boll per plant, boll weight, earliness percentage and lint-yield were investigated.
Findings: The combined variance analysis results for the studied traits showed that the genotype had a significant effect on all traits except for the monopodial length and the sympodial number at the 1% probability level. The comparison of genotype means showed that the genotypes A-NB414 and A-NBK had the highest lint-yield of 5445 and 5154 kg/ha, respectively. The lint-yield of these two genotypes were significantly different from those of the other genotypes. The genotypes A-NB414 and A-NBK also had higher values for plant height, sympodial number, number of boll per plant, and earliness percentage than the other genotypes. The phenotype correlation analysis showed a significant positive correlation between lint-yield and number of boll per plant (r= 0.80**), boll weight (r= 0.36**), and plant height (r= 0.60**) at the 1% probability level. The stepwise regression analysis results showed that the traits of number of boll per plant, boll weight, plant height, and sympodial length had the most significant effects on cotton lint-yield (R2=0.77). According to the results of the path analysis, the number of boll per plant had the most direct effect on lint-yield (P=0.78). The indirect effect of plant height on lint-yield was exerted through the number of boll per plant. The lint-yield of these two genotypes was significantly different from other genotypes. A-NB414 and A-NBK genotypes were superior to other genotypes in plant height, sympodial number, number of boll per plant, and earliness percentage. The phenotypic correlation showed a high and significant correlation at 1 % level of probability respectively between the lint-yield and the number of boll per plant (r= 0.80**), boll weight (r= 0.36**) and plant height (r= 0.60**). The results of step-by-step regression analysis showed that the traits of boll number, boll weight, plant height and sympodial length explain the most lint-yield changes (R2 = 77.59). According to the results of the path analysis, the most direct effect on the lint-yield trait was related to the number of boll per plant (P=0.78). The most indirect effect of plant height was exerted on lint-yield through the number of boll per plant.
Conclusion: Considering the very high correlation between the lint-yield and the number of boll per plant, it can be concluded that the number of boll per plant is the main factor causing genotype differences in lint-yield. Since boll weight, number of boll per plant and plant height have the largest direct effects on cotton lint-yield, these traits can be used as selection criteria. Additionally, since the number of boll per plant has the most direct and indirect effect on lint-yield, it can be used to increase lint-yield in breeding programs.