**4. Discussion**

High genetic variability observed for aflatoxin resistance accumulation was an indication of the presence of novel or favorable alleles for population improvement. Furtherance to this, genotypes identified with reduced aflatoxin accumulation could be exploited in the development of superior hybrids that combine resistance to aflatoxin accumulation and high yields as described previously by Warburton and Williams [6]. The observed significant phenotypic variation for the major, minor and across seasons and locations among the genotypes for aflatoxin accumulation reduction and other agronomic traits suggested that progress could be made in developing well adopted lines with good aflatoxin accumulation resistance.

The significant environmental and genotypic effects detected for aflatoxin accumulation resistance and other agronomic traits indicated variability among the genotypes under different environments. Also the significant genotype x environment interactions observed across the seasons and locations indicates the need for evaluation of genotypes across several environments in order to determine most stable genotypes for aflatoxin accumulation resistance and other agronomic traits. According to Comstock and Moll [15], genotype x environment interactions determined in multi-location trials implied reduced correlation between genotypic and phenotypic values. Zuber [16] identified significant environmental effects on aflatoxin accumulation among commercial hybrids and OPVs in the United States of America. It was not uncommon to observe G x E effects on the genotypes evaluated across contrasting environments in Ghana.

Broad sense heritability among traits ranged from moderate to very high estimates during the major season where adequate rains and less disease pressure was observed. However, significantly lower range of estimates were detected during the minor season (data not shown) confirming earlier reports by [17] who demonstrated significant environmental influence on heritability estimates in cashew. Moderate to high heritability estimates realized in this study, suggested that possible gains are achievable in hybrid maize development for high yields and aflatoxin accumulation resistance.

Further evaluation of the agronomic traits revealed a range of genotypic influence on several parameters studied. Significant among them were days to 50% pollen and silking which clearly categorized the genotypes into the three well defined and established classes of extra early, intermediate and late types as reported by Badu- Apraku [18] who categorized maize genotypes into the different maturity groups. Such information is critical and necessary to guide planting periods in breeding nurseries designed to cross among the maturity groups for trait introgression and further improvement for hybrid development.

The combined analysis showed that most genotypes belonged to the intermediate group while a few were extra early or late maturing. For instance, genotype MP715 has been reported in previous studies [19] as highly resistant to aflatoxin accumulation with delayed silking ranging between 70 and 80 days when grown in a temperate environment, however in this study it ranged between 62 and 67 days under tropical conditions in Ghana, where it is evidently adapted. On the other

hand, genotypes identified with delayed silking dates included MP719, TZI8, NC334, OBAATANPA, CML176, CML247, CML287, CML343 and CML5.

Aflatoxin accumulation levels during the major season were comparatively lower across environments although levels in Wenchi were slightly higher. Unlike the major season, the accumulation levels in the minor season were considerably higher across environments with Wenchi still ranking highest. This observation agrees with the findings of several authors [2, 6, 20, 21] who reported the existence of positive correlation of drought and heat on aflatoxin accumulation level. The Guinea savanna transition environment, which appears to be relatively drier, was conducive for aflatoxin production.

The ranking of top ten resistant genotypes across environments revealed a consistent set of genotypes (although in different ranking order of resistance per environment) during both major and minor seasons. Genotypes which consistently displayed stable resistance across the environments included MP715, NC298, MP705, MP719, CML287 and TZEEI- 24 while the rest appeared less stable. Two local extra early lines (TZEEI-24 and TZEEI-4) were identified as sources of potential resistance to aflatoxin but their level of resistance was not as good as the Mississippi lines specifically bred for resistance and that further evaluations of these two locally adapted lines may be required to confirm their levels of resistance to aflatoxin accumulation.

It was also obvious from this study that, majority of the inbred lines outperformed the OPVs and the populations' in-terms of reduced aflatoxin accumulation levels which was in agreement with the previous findings of Zuber [16] who reported superiority of hybrids (inbred combinations) over OPVs in-terms of measured aflatoxin accumulation resistance across locations and years in the United States.

The levels of total aflatoxin accumulated by the resistant genotypes in this study are comparable to the levels previously reported by William and Windham [5] and Brown [2] where a set of hybrids were evaluated for aflatoxin accumulation resistance. Information obtained from the total aflatoxin accumulation levels among the genotypes could therefore guide the selection of appropriate parental candidates for future aflatoxin resistance breeding in Ghana. Breeding for resistance involves several approaches of which trait correlations is paramount. Indirect selection of one trait simultaneously improves other traits that are significantly correlated.

Zuber [16] discovered strong significant correlation between insect damage and aflatoxin accumulation during evaluation of OPVs and released commercial hybrids in the United States of America. The correlation observed between insect damage and aflatoxin accumulation in this study was similar to that of Zuber [16] as well as the observations made by Williams [22] and Ni [23]. Significant positive correlations were also observed between aflatoxin accumulation levels and plant aspect and cob aspect whilst ears per plant showed a weak negative correlation. Stay-green and open- tip did not correlate with aflatoxin accumulation levels as recounted in other studies [24, 25].

The study of genetic relationship among genotypes which was based on significant agronomic traits assigned all genotypes into three main groups when 40% of the variation among the genotypes was explained. On the contrary, only two main groups were realized when 20% of the genotypic variation was explained by the significant traits used for the grouping. In the case of the two groups, all exotic genotypes were assigned into one group except Ki3, CML108 and NC340 which clustered together with the local genotypes. This observation was not surprising since most of the local genotypes were sourced from CIMMYT and may have similar pedigree or ancestry records.

It appears that the top resistant genotypes which also clustered in one common group perhaps originated from a common ancestry of Tuxpe˜no germplasm native to Mexico which exhibits tropical characteristics coupled with aflatoxin accumulation resistance [6].

Although the analysis of the multiple phenotypic traits assigned genotypes into distinct groups, it showed a low corroboration when compared to other molecular methods (data not shown). This is probably because the expression of most agronomic traits are influenced by the environment.
