Analysis of Correlation,Cluster and Variation on the Main Agronomic Characters of the Upland Cotton Cultivars or Lines

[Objective]The present research was carried out to effectively provide the useful information and academic base for choosing and breeding high yield cotton variety in Xinjiang.[Method]The correlation,cluster and variation of the main agronomic characters were analyzed in 29 cotton cultivars or lines by using multivariate statistics.[Result]All the 29 cotton cultivars or lines were clustered into 2 major groups or 4 small categories according to the main characters.The phenotypic characteristics of those subtypes were obviously different.In all the main characters,variation coefficient of first fruit node,fruit branches,plants per 667 m~2 and plant height were the most group,those of the characters of yield were the middle,and those of the characters of fiber quality were the least.Lint yield of cotton had a positive correlation with seven characteristics including bolls per plant,plants per 667 m~2,lint percentage,first fruit node,weight per boll, fruit branch and fiber length,but a negative correlation with seven characteristics including plant height,seed and lint index,fiber strength,uniformity,elongation rate and micronaire.Most importantly,correlation analysis indicated that lint yield was positively relative to plants per 667 m~2 and lint percentage at 0.05 level. [Conclusion]Lint percentage and plants per 667 m~2 are the main effecting factors of lint yield,but bolls per plant and weight per boll are the second effecting factors.Consequently,augmenting plant density,choosing the upland cotton cultivars or lines of the higher lint percentage are one of important approaches for increasing lint yield of the upland cotton in Southern Xinjiang.