SCREENING COTTON FOR TOLERANCE TO HOPLOLAIMUS-COLUMBUS

In order to adequately plan experiments to measure tolerance of cotton to Hoplolaimus columbus, coefficients of variation (CVs) and error va riances of yield were examined to determine the minimum number of repl icates and environments needed to measure specified differences for to lerance index (TI) and field tolerance. In two greenhouse trials with 84 genotypes, the error variance was 1,423 and would have required 231 replicates to ensure detecting a 10% genotypic difference in TI. In t he 3-yr, one-location field study reported, CVs ranged from 10.0 to 25 .4%, while error variances for TI ranged from 68 to 512. The pooled er ror variance (243) was used in calculating the minimum number of repli cates and environments. There was no genotype X environment interactio n for TI. To detect a 20% difference in TI, a minimum of 14 replicates in one environment or six replicates in two environments was required . Field tolerance, on the other hand, revealed a highly significant ge notype X environment interaction and required as many as 23 environmen ts with four replicates per environment to detect a 10% difference at 0.05. Choosing an cu level of 0.20 instead of 0.05 would reduce the mi nimum number of replicates and/or environments by over one-half. This information should be useful in planning experiments to screen cotton germ plasm for tolerance to Hoplolaimus columbus in the future.