Genetic Divergence and Importance of Traits in Luffa cylindrica L.

This work aimed at estimating the genetic divergence between luffa accessions (Luffa cylindrica (L.) M. J. Roem) by means of multivariate analysis techniques and at verifying the importance of the assessed traits for the study of genetic divergence. Two trials were carried out in the municipality of Mossoró, in the state of Rio Grande do Norte, Brazil. The trials were carried out in incomplete block design, comprising two experiments in completely randomised blocks with four regular treatments and two treatments common to both experiments, amounting to a number of ten treatments, with three replications. Each luffa accession composed one treatment. Eleven descriptors (three quantitative and eight qualitative) were measured in order to quantify genetic divergence, using Gower’s algorithm as a measure of dissimilarity, and one joint analysis was carried out considering quantitative and qualitative descriptors simultaneously. The UPGMA hierarchical method was adopted to group the genotypes, and the Singh criterion was elected in order to quantify the contribution of the traits to genetic divergence. It could be verified that there are accessions presenting great divergence from one another. Accessions BUCHA06 and BUCHA22 were the most divergent from the others, and accessions BUCHA06 and BUCHA10 were noted to be the most similar. The median portion fiber texture, the upper portion fiber texture, and fruit mass were found to be the most important traits for estimating genetic divergence.