STABILITY PARAMETERS PROPOSED BY LIN AND BINNS (1988) COMPARED WITH REGRESSION METHOD

The stability parameters proposed by Lin & Binns (1988) method for yie ld of seed cotton were compared with the regression method proposed by Eberhart & Russell (1966). The experiments were composed by nine cult ivars tested at randomized complete block design with six replicates c arried out by CNPA in sixteen sites on Northeast Brazil in 1991 and 19 92. The estimate comparison between the two methods was performed by S pearman rank correlation. The cultivars CNPA 6H and CNPA 85-241 showed high yields, beta(1) greater than or equal to 1, R(2) > 95%, low valu es of P-i and contributed only with 2.1 and 3.5% of total interaction while the CNPA Giorge I was the more instable. The Spearman rank corre lation between beta(1) and P-i in 1992 was -0.96 significative by t te st (P less than or equal to 0.01) indicating that these parameters pro vided the same results. The advantage of P-i utilization is that there is no restriction on the independence of variables and also it indica tes the materials with performance near the maximum value in different environments tested.