tatistic and experimental designFor both two experiments, treatments were distributed in a randomized block design in a factorial scheme(3 · 3 · 2) + 1 with four replicates. The first factor was the glyphosate rates (800, 1200 and 2400 g a.e. ha)1), the second was the growth stage (V2, V4 and V6) and the third factor was cultivar representing the ‘first and second generation’ cultivars (cv. BRS 242, GR1 and cv. AG3539,GR2, respectively). The additional treatment was a nonglyphosate treated.The data errors passed the normality test of Shapiro and Wilk (1965), and because there was homogeneity of error variances, the data for the two repeated experiments were combined. No transformations were necessary.Data were subjected to analysis of variance, and when F values were significant (P <0Æ01), to regression analysis. Data were analysed using proc mixed by sas statisticalprogram (SAS Institute, Cary, NC), and equations were adjusted using the polynomial model^y ¼ a þ bx þ cx05by SigmaPlot 10.0 statistical package (Systat Software,SPSS, Chicago, IL).
การแปล กรุณารอสักครู่..
