Understanding the nature and magnitude of variability existing among sweetpotato genetic materials for important traits is vital for the effective utilization of such materials for breeding purposes. Eighteen landraces from diverse origins, plus three released cultivars as checks, were evaluated in two contrasting locations, using nine agronomic and eight morphological traits of the crop, to estimate the nature and magnitude of the variability among the genetic materials, to determine the relationships among the traits, and to identify the important yield-related traits among the collections using multivariate tools. The principal component analysis identified number of marketable and unmarketable roots, total number of roots, weight of marketable and unmarketable roots, total root weights, incidence and severity of rootCylas spp., length of biggest, medium and smallest marketable roots, number of branches, as well as stand count at harvest as important traits that could be used to differentiate the landraces. The canonical variate analysis showed that the observed variation among the traits occurred mostly between-groups than within-groups, and that it was largely influenced by total root weight, weight of marketable roots, number of marketable roots, and total number of roots. Generally, all the traits, except stand count at harvest, exhibited positive and significant (P<0.01 and P<0.001) correlation with total root weight (yield). Most of the traits also exhibited significant relationships among them. However, the use of forward selection multiple regression analysis revealed weight of marketable and unmarketable roots, as well as total number of roots as the most important yield component traits that could be used to improve sweetpotato. Thus, our work identified the existence of inherent variability in the local germplasm collections, and the traits that could be used to exploit the observed variability, eliciting important relationships among the traits in the process.
Key words: Sweetpotato, genetic variability, multivariate analyses, yield components.
Copyright © 2019 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0