Comparação e seleção de populações de trigo: componentes de rendimento, qualidade industrial e resistência à giberela

Abstract

Being an important carbohydrate source, wheat is a crop of great relevance in human food. Brazil is not self sufficient regarding this crop, the imports accounting for half of the national consumption. This lack of wheat in the country is mainly due to climatic instability in the main producing states, besides, commercial factors, institutional variables and the lack of attractive public policies to the wheat farmers. Many breeding programs are carrying out investigations to obtain higher yielding genotypes that present good technological quality, with resistance to the main diseases, stable and adapted to the growing regions. However, all these attributes are governed by quantitative traits, making it difficult to select and obtain superior genotypes. The characterization of the genetic variablility of a population, the prediction of the truly inheritable portion and the gain by selection of the target character become promising tools and of great value to the breeding programs. The main objective of this work was the comparison of wheat populations for yield components, industrial quality and resistance to gibberella. The populations originated from the following crosses: 1: Abalone x Fundacep Nova Era; 2: Ônix x Fundacep Raízes; 3: CD 104 x Fundacep Cristalino; 4: Fundacep Cristalino x Fundacep Nova Era; 5: CD 104 x Fundacep Raízes and 6: Fundacep Nova Era x CD 120, in three distinct generations (F4, F5 and F6). The traits of spike length (SL), mass spike (MS), number of grains per spike (NGS), mass of grains per spike (MGS), grain yield (GY) industrial quality characters and severity infection and gibberella infected grains. The use of methods for stability and adaptability analyses allowed the identification of superior genotypes for environments in favorable and unfavorable conditions. However, the Lin and Binns method and the mixed model method (BLUPs) present redundant information. The best method to identify the most stable genotypes and also the most productive was the mixed model method (BLUPs). The inheritable portion of the phenotypic variance was very small, showing a large effect of the environment. The heritability among families was higher, being indicated to carry out selection among the families to obtain larger gains. Population 6 presented the highest gains for SL, MS and NGS. Eleven families (populations 2, 4 and 6) were selected to simultaneously increase the four characters evaluated. The populations showed variability for the traits related to industrial quality, as well as the negative association between grain yield and protein content. The physical, chemical and rheological tests for industrial quality determination were correlated. Population 4 stands out for the industrial quality characters evaluated. The family 49 is promising for use in cross-blocks to increase protein content. The cultivars Fundacep Nova Era and Fundacep Cristalino showed lower severity of gibberella infection and number of gibberella infected grains, respectively, and was not surpassed by any population. Population 4 is less susceptible to gibberella infection, especially families 25, 104, 106 and 107.