CROP PERFORMANCE TRIALS UNDER IRRIGATED AND DRYLAND CONDITIONS

Corn (Zea mays L.) and soybean [Glycine mar (L.) Merr.] performance tr ials are sometimes conducted under both irrigated and dryland conditio ns in the southeastern USA. It is not known whether irrigated trials a re better trials statistically or if irrigation causes relative rank c hanges in hybrid or cultivar performance. To answer these questions, c orn data were examined from four different environments and three matu rity groups, and soybean data were examined from five different enviro nments and three maturity groups between 1987 and 1990. Water manageme nt x hybrid/cultivar interactions were evident in 8 of 12 corn data se ts and 7 of 16 soybean data sets. Combining data across years eliminat ed the water management x hybrid/cultivar interactions in all soybean data sets and five of nine corn data sets. Irrigated corn trials produ ced year x hybrid interactions in seven of nine data sets while drylan d trials produced year x hybrid interactions in four of nine data sets . Irrigated trials were only slightly better than dryland trials in ac counting for variation; i.e., larger r2 values but produced smaller er ror variances in four of 12 data sets of corn. Irrigated trials separa ted entries better in nearly half of the soybean data sets. The data s uggested that irrigated trials were not statistically better trials th an dryland trials in all situations. Water management x hybrid/ cultiv ar interactions suggest that data for irrigated and dryland trials' sh ould be reported separately, and the year x hybrid/cultivar interactio ns indicated that multi-year data should be examined during hybrid/ cu ltivar selection. Error variances, F probability values, and r2 values should be examined to determine level of precision when comparing tri als of equal composition.