It was found that there were differences in each statistics of
t-test, i.e., in standard error of the location-specific effect estimate,
degrees of freedom, t-value and hence p-value in testing the
location-specific genotype effect difference (results not shown). For
brevity, only the discrepancy number of significant (˛= 0.05) pairs
between the separate ANOVA and the BLUP in the difference test
about location-specific genotype effects is summarized in Table 6.
It showed that there was number discrepancy of significant genotype
pairs between the two analysis methods at all locations for
five trials (the trials of group1 of zone 2, group 1, 2, 3 and 4 of zone
3) and at most locations for the other nine trials.