Abstract The genotype x environment interaction is frequently observed in many crops and studies on environmental stratification and genotype adaptability have been proposed to understand it. The aim of this study was to carry out factor analysis in data from multi-environment experiments by the mixed model approach (REML/BLUP). Instead of adjusted phenotypic means, a matrix containing the genotypic effects added to the effects of the genotype x environment interaction (G+GE) was used, predicted via REML/BLUP in joint analysis (designated as R-FGGE). In the study, data from 36 common bean lines evaluated in 15 environments were used. By this proposal, 46.7% of the environments were gathered in two groups, one with four and the other with three environments. The R-FGGE has the same characteristics as the previous proposals, that is, ease of identification of mega-environments and genotypes with broad adaptability, along with the advantages associated with the mixed model methodology.

A cobertura vegetal do solo é decisiva para redução dos efeitos erosivos do impacto direto das gotas de chuva na superfície do solo. Desta forma, objetivou-se com este estudo determinar o índice de cobertura vegetal (CV) e desenvolver modelos para sua estimativa para a cultura da soja, usando os atributos climáticos no período de chuvas intensas no Sul de Minas Gerais. As determinações da CV foram feitas semanalmente, na área experimental do Departamento de Ciência do Solo, Universidade Federal de Lavras, no período de novembro de 1999 a maio de 2000, em 28 cultivares de soja com potencial para cultivo nesta região. Para avaliação da cobertura vegetal foi utilizada a metodologia descrita por Stocking (1988). Na modelagem procurou-se relacionar a CV com os valores acumulados dos seguintes atributos climáticos: temperatura média (Tmed), precipitação (PREC) e umidade relativa do ar (UR). Os valores de cobertura vegetal apresentaram uma amplitude de variação de 56 a 83%, sendo BR 162, LO 12 L e M. Soy 108 as cultivares mais eficientes e FT Abyara e Tucano as menos eficientes. O hábito diferencial de crescimento das cultivares ajuda a explicar esse comportamento. O modelo ajustado adequado para estimativa da CV foi: CV = 116589,976 + 0,422 . Tmed + 0,132 . PREC - 0,095 . UR + 0,000024 . Tmed², R² = 0,99 (P < 0,01). A determinação da CV nas diferentes fases de desenvolvimento da cultura é de grande importância, uma vez que seu estabelecimento coincide com o período de maior potencial erosivo das chuvas na região estudada.

Vegetal cover of soil is decisive to reduce the erosive effects of direct impact of raindrops on the soil surface. Therefore, the objective of this study was to determine the vegetal cover (CC) index and to develop models to estimate it for soybean cultivars, using climatic attributes in the period of intense rains in the South of the State of Minas Gerais in Brazil. CC was measured weekly in the experimental area of the Department of Soil Science, Federal University of Lavras, from November 1999 to May 2000, for 28 soybean cultivars with yield potential in this region. To evaluate the vegetal cover, the method described by Stocking (1988) was used. In the modeling, CC was related with the accumulated values of following climatic attributes: medium temperature (Tmed), precipitation (PREC), and relative humidity of the air (RH). Vegetal cover values presented an amplitude from 56 to 83%, being BR 162, LO 12 L and M. Soy 108, the more efficient cultivars, and the FT Abyara and Tucano, the least efficient ones. The differential growth habit of the cultivars helps to explain this behavior. The best adjusted model for the estimative of CC was: CC = 116589.976 + 0.422, Tmed + 0.132, PREC - 0.095, RH + 0.000024 Tmed², R² = 0.99 (P < 0.01). The knowledge of CC for different development phases of the crop is of great importance, taking into account that its establishment time coincides with the highest erosive potential of the rains in the studied region.