engleski

OPTIMIZING MAIZE HYBRID TESTING IN CROATIA

In Croatia, maize hybrids were tested by the Registration office for three years at four locations (FAO 200 and 600) or six locations (FAO 300, 400, 500) in four replications. However, no published data existed on whether all locations are equally suitable for hybrid testing. Optimal site/location should be representative possessing high correlation between site mean and total mean across genotypes and reliable measured by site specific heritability characterizing the accuracy of the experiment. The objectives of this study are i) to determine differences of several locations in Croatia according to their suitability for yield testing and ii) to recommend some methodological corrections for setting of maize yield trials in Croatia. Our experiment consists of complete data sets of 8 locations in three years with 24 genotypes and four replications respectively. The yield trials were conducted at the locations of Northern Croatia (major maize growing region), with a higher density of sites in the most intense maize growing area of North-East Croatia. Each of the 24 location/year combination was considered as a random (macro)-environment. The relationships among mean grain yields and correlations between environmental mean and total mean for each location averaged over 24 maize hybrids varied considerably, indicating the differences in suitability for selecting genotypes among locations. Some of the high yielding sites show consistent tight correlation with yield over the range of environments. Some of the best yielding sites had consistent high site specific heritability estimates indicating constant high accuracy of trials. Due to differences of locations in Croatia in their suitability for yield testing, there is a necessity for weighing sites or excluding sites with extreme low estimates of site specific correlations and heritabilities. According to a simulation of decreasing number of replications to optimize the testing process, it seems appropriate to reduce the replication number to three with no substantial reduction of site specific heritability estimates.

maize; hybrids; allocation

nije evidentirano