Selection of elite Rhizobium strains by biometric techniques for inoculation in common bean

Abstract

Variation in plant responses to inoculation with rhizobia strains in different environments can be analyzed with the aid of biometric techniques. The Annicchiarico methodology and the Additive Main Effects and Multiplicative Interaction Analysis (AMMI) and Toler models are here applied to the genotype × environment (GE) interaction to determine the adaptability and phenotypic stability of N2-fixing bacteria in symbiosis with common bean (Phaseolus vulgaris L.) in environments with high, moderate, and low acidity and with deficiencies or excessive concentrations of nutrients. A randomized block design was used in eight environments, with four replications and seven bacterial genotypes, including two controls with native soil bacterial communities (with or addition without mineral-N fertilization). Individual and combined analyses of variance and biometric analyses were conducted on the yield values. CIAT899, UFLA02-127, and native rhizobia were the most stable genotypes in relation to environments (AMMI models). CIAT899 and native rhizobia are the least adapted (Toler models), whereas the genotype UFLA02-127 has the best response to the environments (Toler models). The Annicchiarico method showed that the reliability of adoption of UFLA02-127 and native rhizobia fertilized with mineral N exceeds 85%, indicating high probability of increase in common bean yield if adopted in the field. Conversely, native rhizobia have <65% reliability.