Morpho−Agronomical Characterization of Taro (Colocasia esculenta) Accessions in Ethiopia

One hundred accessions of taro (Colocasia esculenta) were tested at Jimma Agricultural Research Center dur-ing the 2011 growing season by using 10 x10 simple lattice design with two replications. The objectives of the study were to assess the nature and extent of diversity with in collected accessions based on key morphological descriptors and evaluate the accessions based on yield and yield related traits. Data on 17 qualitative and 13 quantitative traits were measured. Both sets of data collected were subjected to multivariate analysis using principal component (PCA), cluster analysis and of analysis of variance to determine the variability among accessions. The results of PCA based on qualitative traits reveled that characters have large contribution to the variability. The two-dimensional plot of the first two PCs showed a separation between accessions of big sized. Cluster analysis based on qualitative characters indicated the formation of seven clusters and existence of variability, based on their foliar and subterranean traits. Shannon-Weaver diversity index (H=0.27) showed that low levels of diversity existed among 100 C. esculenta based on the frequency of phenotypic characters that were considered. The analysis of variance for quantitative traits showed that number of active leaves/plant, petiole length and maximum horizontal distance highly significance differences exists for majority of characters studied and also path coefficient analysis at genotypic level showed that there are a number of characters exerting a maximum direct positive effect on fresh tuber yield suggesting that these characters are good contribution to the fresh root yield in taro. Overall, this study indicated that taro accessions collected from south and southwestern Ethiopia have an enormous wealth of traits variation for qualitative and quantitative characters, indicating huge potential for its genetic improvements through selection and hybridization. Therefore, selections based on these characters are efficient to maximize the fresh root yield and the genetic improvement of the crop.