Abstract: Screening rape varieties suitable for planting in Hebei low plain area may allow ecological and economic benefits simultaneously achieved,which may provide theoretical basis for the selection of rape varieties and the adjustment of planting structure in similar climate areas. Through the analysis of variance,correlation and cluster,the variations in agronomic traits of 16 rape varieties were analyzed. There were the most obvious differences in branching parts among varieties,with the variation coefficient of 61.82%,the lowest branching parts for variety JR5（1.2 cm）,and the highest branches for variety 14T×38（15.9 cm）.The second significant difference was in yield and secondary branching number,with variation coefficients of 26.69% and 24.89%,respectively. The yield of variety 14T×38 was the highest,and that of the 20SY-13 was the lowest. The varieties with the largest and smallest number of secondary branches were 20SY-13 and Tianyou 142,respectively. The correlation between rapeseed yield and a number of primary branches and the main inflorescence pod number were very significantly positive（P<0.01）,plant height was significantly positive correlated with branch height,length of main inflorescence and the number of main inflorescence pod,and a number of primary branches and the main inflorescence pod number was significantly positive correlated（P<0.05）,while secondary branch number was very significantly negative correlation with the main inflorescence pod number. When the Euclidean distance was 5,16 varieties can be grouped into 5 categories. The higher yield could be obtained while cultivars with higher plant height,more primary branches and pods in the main inflorescence,less secondary branches and normal overwintering were selected for planting in the local area. The safely overwintering should be considered if 20SY-13 is planted in autumn in this area.

Key words: winter rape, agronomic trait, variance analysis, correlation analysis, clustering analysis