Diversity among elephant grass genotypes using Bayesian multi-trait model.

Elephant grass is a perennial tropical grass with great potential for energy generation from biomass. The objective of this study was to estimate the genetic diversity among elephant grass accessions based on morpho-agronomic and biomass quality traits and to identify promising genotypes for obtaining hybrids with high energetic biomass production capacity. The experiment was installed at experimental area of the State Agricultural College Antônio Sarlo, in Campos dos Goytacazes. Fifty-two elephant grass genotypes were evaluated in a randomized block design with two replicates. Components of variance and the genotypic means were obtained using a Bayesian multi-trait model. We considered 350,000 iterations in the Gibbs sampler algorithm for each parameter adopted, with a warm-up period (burn-in) of 50,000 Iterations. For obtaining an uncorrelated sample, we considered five iterations (thinning) as a spacing between sampled points, which resulted in a final sample size 60,000. Subsequently, the Mahalanobis distance between each pair of genotypes was estimated. Estimates of genotypic variance indicated a favorable condition for gains in all traits. Elephant grass accessions presented greater variability for biomass quality traits, for which three groups were formed, while for the agronomic traits, two groups were formed. Crosses between Mercker Pinda México x Mercker 86-México, Mercker Pinda México x Turrialba, and Mercker 86-México x Taiwan A-25 can be carried out for obtaining elephant grass hybrids for energy purposes.

Stability and adaptability of grain sorghum hybrids in the off-season.

We analyzed productivity data obtained from experiments on grain sorghum conducted in 7 locations of its cultivation in Brazil. A total of 25 hybrids were analyzed, of which 22 were pre-commercial and 3 were cultivars. The Wricke and Purchase et al. methods were highly consistent in identifying individuals with low contributions to genotype x environment interactions. The Lin and Binns method proved to be easily applicable and interpretable but it was not efficient in detecting individuals with specific adaptations. An additive main effect and multiplicative interaction (AMMI) model indicated the suitability of cultivar 1G282 for the cities of Guaíra, Sete Lagoas, and Vilhena, and hybrids 0307087 and 0307091 for the southeast of Goiás. The associations of the Eberhart and Russell method with AMMI indicated that 0307071, 0307131, 0307511, and 0307651 showed adaptability to favorable environments. Hybrid 0009061 stood out as the most adaptable and stable cultivar.