Avaliação da eficiência da rede experimental da Embrapa em dez anos do programa de melhoramento do feijoeiro-comum
Nenhuma Miniatura disponível
Data
2016-02-25
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Goiás
Resumo
The assay of VCU (Value of Cutivation and Use), called of final trials, are in
network conducted and are systematized in an organized manner which includes the states
responsible for over 90% of national production of common bean. The final trials are
realized in different locations, seasons and years, aiming of superiors line selection:
productive, stable, adapted and with disered agronomic attributes. The genotypes
environments interaction (GE) in plant breeding program has implications numberless,
mainly in phase of lines final evaluation. In this case, alternatives to minimize the effect of
the GE interaction should search. Among them stand out: identification of predictable
behaviors of the cultivars, stratification of recommendation regions in homogeneous subregions and decomposition of GE interaction, aiming at identifying of the factors whose
interaction is more expressive. The aim of this study was realize environmental
stratification of network final trials and decompose the genotypes interactions with
environment factors (locations, seasons and years), during ten years of common bean
program at Embrapa of grains "carioca" and black, in main three producing regions in
brazilian (R1-South-Central, R2-Central, R3-Northeast). In study of environmental
stratification also was evaluated representativeness of base locations of this program. Data
for grain yield of the conducted trials in random blocks, with three replications, during
2001 until 2012, 670 trials total (292 in R1, 250 R2 and 128 R3). During all period were
evaluated 88 genotypes of “carioca” group (76 lines and 12 checks) and evaluated 71
genotypes in black group (58 lines and 13 checks). In all regions the local effect were
considered as random and genotypes fixed. Environmental stratification was separately
realized in the main crop seasons of each region, utilized factors analysis methodology and
ecovalence. For simultaneously utilization were estimated redundance index joint:
redundance percentage (RP), redundance fraction (RF) and redundance percentage in
season (RPS). The elimination local was established based observing the three index
redundancy values (RP ≥ 15%, RF ≥ ½ e RPS ≥ 50%). To R1, Candoi-PR, Palma-SC,
Catanduvas-PR, Campos Novos-SC, Laranjeiras do Sul-PR, Roncador-PR e Dois
Vizinhos-PR in rainy seasons and Passo Fundo-RS, Prudentópolis-PR e Dourados-MS in
dry seasons were considered redundant, should be eliminated of the trials network. The
new evaluation network would be formed by 29 municipalities in rainy and twelve
municipalities in dry, considered informative at final trials network. In R2, Morrinhos-GO,
Inhumas-GO, Urutaí-GO, Santo Antônio de Goiás-GO, Cristalina-GO e Planaltina-GO e
Sinop-MT in rainy and Porangatu-GO, Goiatuba-GO, Urutaí-GO, Senador Canedo-GO,
Cáceres-MT, Primavera do Leste-MT e Brasília-DF in dry were redundantes and should be
eliminated of the network. This region stayed with 12 municipalities available for
realization of the rainy season assays and ten in winter season. To R3, should be eliminated
of the network Coronel João Sá-BA, Poço Redondo-SE, Nossa Senhora das Dores-SE,
Simão Dias-SE, because, no aggregate in informations to evaluation of the lines, it was
uninformative. For this region remained 14 municipalities to composition of the network.
Can be concluded than evaluation network at Embrapa represented efficient form variation
of the cultive with common bean in Brazil and than environments base of the breeding
program in regions South-Central and Central, are consistentes and informative, except
during in the rainy season in that Santo Antônio de Goiás-GO is not mandatory enter the
assays network. Especially, Ponta Grossa and Santo Antonio de Goias-GO, than are
evaluation strategic points of the segregation populations and initial lines, of the common
bean breeding program genetic at Brazil. To check with which environmental factors
(locations, seasons and/or years) of the common bean genotypes interaction was more
expressive, the GE interaction was decomposed in genotypes x years, genotypes x seasons
and genotypes x locations. Were utilized grain yield of data 501 trials, 207 in R1, 202 in
R2 and 92 in R3. For the fisrt two regions the analysis were partially balanced, this is, only
utilized trials data of the locations had evaluation at least one season and in two years each
cycle. In R3 the joint analyses were balanced, considered just assays data of the locations
with evaluation in two years of the each cycle. In all regions genotype effects, local, season
and year were considered fixed. To identify importance of the each variation source from
joint variance analyse was estimated contribution of each in relation overall variation, used
estimated determination coefficient (R2). The contribuition average over years for
genotypes effect in all regions, at “carioca” and black group was important less, followed
followed year and season. However, local with interaction was more pronounced in all
regions. The interaction GxL was more important in the three regions for both groups,
compared with intractions GxS and GxY. Thus, in the R1, R2 and R3, priority should be
given of locations increment number evaluation, over seasons and years, aiming at more
efficient and safe recommendation of new cultivars to the main producer regions common
bean of Brazil.
Descrição
Citação
PONTES JÚNIOR, V. A. Avaliação da eficiência da rede experimental da Embrapa em dez anos do programa de melhoramento do feijoeiro-comum. 2016. 253 f. Tese (doutorado em Agronomia) - Universidade Federal de Goiás, Goiânia, 2016.