Yield Stability of Maize Hybrids Evaluated in National Maize Cultivar Regional Trials in Southwestern China Using Parametric Methods

Assessment of yield stability is an important issue for maize （Zea mays L.） cultivar evaluation and recommendation. Many parametric procedures are available for stability analysis, each of them allowing for different interpretations. The objective of the present study was to assess yield stability of maize hybrids evaluated in the National Maize Cultivar Regional Trials in southwestern China using 20 parametric stability statistics proposed by various authors at different times, and to investigate their interrelationships. Two yield datasets were obtained from the 2003 and 2004 national maize cultivar regional trials in southwestern China. A combined analysis of variance, stability statistics, and rank correlations among these stability statistics were determined. Effects of location, cultivar, and cultivar by location interaction were highly significant （P〈0.01）. Different stability statistics were used to determine the stability of the studied cultivars. Cultivar mean yield （Y） was significantly correlated to the Lin and Binns stability statistic （LP, r=0.98^＊＊ and 0.97^＊＊ for 2003 and 2004 trials, respectively） and desirability index （HD, r=0.38 and 0.84^＊＊ for the 2003 and 2004 trials, respectively）. The statistics LP and HD would be useful for simultaneously selecting for high yield and stability. Based on a principal component analysis, the parametric stability statistics grouped as four distinct classes that corresponded to different agronomic and biological concepts of stability.

Yield Stability of Maize Hybrids Evaluated in Maize Regional Trials in Southwestern China Using Nonparametric Methods

Analysis of multi-environment trials （METs） of crops for the evaluation and recommendation of varieties is an important issue in plant breeding research. Evaluating on the both stability of performance and high yield is essential in MET analyses. The objective of the present investigation was to compare 11 nonparametric stability statistics and apply nonparametric tests for genotype-by-environment interaction （GEI） to 14 maize （Zea mays L.） genotypes grown at 25 locations in southwestern China during 2005. Results of nonparametric tests of GEl and a combined ANOVA across locations showed that both crossover and noncrossover GEI, and genotypes varied highly significantly for yield. The results of principal component analysis, correlation analysis of nonparametric statistics, and yield indicated the nonparametric statistics grouped as four distinct classes that corresponded to different agronomic and biological concepts of stability. Furthermore, high values of TOP and low values of rank-sum were associated with high mean yield, but the other nonparametric statistics were not positively correlated with mean yield. Therefore, only rank-sum and TOP methods would be useful for simultaneously selection for high yield and stability. These two statistics recommended JY686 and HX168 as desirable and ND108, CM12, CN36, and NK6661 as undesirable genotypes.